Mapping the policy networks

Material Information

Mapping the policy networks a case study of the Korean foreign labor policy
Kim, Young-Jung
Publication Date:
Physical Description:
xvii, 302 leaves : ; 28 cm


Subjects / Keywords:
Foreign workers -- Government policy -- Korea (South) ( lcsh )
Foreign workers -- Government policy ( fast )
Korea (South) ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 284-301).
General Note:
School of Public Affairs
Statement of Responsibility:
by Young-Jung Kim.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
656383096 ( OCLC )
LD1193.P86 2010d K55 ( lcc )

Full Text
Young-Jung Kim
B.A., Seoul National University, 1992
M.P.A., Seoul National University, 1994
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Public Affairs

This thesis for the Doctor of Philosophy
degree by
Young-Jung Kim
has been approved
Peter deLeon
Linda deLeon
Chul-Young Roh
U foe

between policy actors (PPP) play a role as the glue of policy coalitions in policy
networks, and the PPP is the most influential factor among them; (2) policy actors
deep core beliefs are not cogently related to the other belief levels, such as policy
core beliefs and the secondary aspects of beliefs, indicating that the hierarchical
structure of policy actors belief systems proposed by the Advocacy Coalition
Framework may be rejected; (3) collective action between policy actors is more
influenced by the trust level between policy actors than any other variable, and the
effects of the PBS, the PPI, and the PPP upon communication density mostly
disappeared when trust level between policy actors was controlled; (4) the PPP is
the most significant variable as an antecedent to trust and trustors propensity to
trust, and the PBS is the other significant variables in the analysis; and (5) policy
actors influence is significantly influenced by policy actors trustworthiness and
their centrality in the network among a variety of variables, and policy actors
trustworthiness is more highly associated with their influence level than their
This abstract accurately represents the content of the candidates thesis. I
recommend its publication.
Peter deLeon

My special thanks to my academic advisor, Peter deLeon, for his constant support
and encouragement for my research. Without his deep concern and detailed
editorial feedback throughout the process, this thesis would not have been
completed. While I admit responsibility for any remaining faults in this thesis, Mr.
deLeon must be acknowledged for significantly enhancing its quality.
I also owe much to the members of my committee: Donald Klingner, Linda deLeon,
Paul Teske, and Chul-Young Roh. They helped me adapt myself to the degree
course and develop my thesis. Professors Donald Klingner and Linda deLeon were
the first people I met in this degree course. They paid particular interest in my
difficulties as a foreign student and encouraged me to proceed on my academic
journey in the U.S. Professor Chul-Young Roh gave me much advice in developing
my thesis based on his own experience as a foreign student. His careful guidance
was another blessing for me. Paul Teske, Dean of the School of Public Affairs, also
gave me much help in advancing my thesis, particularly in developing my research
I would like to thank other faculty and staff members of the School of Public
Affairs. The faculty members significantly contributed to my academic experience
with their passionate lectures and constant interest in providing academic
opportunities. Also, the staff members continuously helped me proceed in my
degree course without encountering any administrative problems.
I appreciate the Korean Ministry of Labor for supporting my study in the U.S. As a
Korean government official, I received the opportunity to study public policies
abroad, which was invaluable in deepening my experience and knowledge about
public policies. This opportunity, I believe, will contribute to enhancing my ability
to designand handle public policies as a practitioner.
Lastly, I would thank my wife, Dr. In-Moon Kim, and two daughters, Mi-Sung and
Min-Jung, for their patience and support throughout the process. They were the
source of my power to continue and to complete this long academic journey. With
their love, I could concentrate upon advancing this thesis.

The Third Round (2000 2003): Adoption of the Employment
Permit System for Foreigners...........................34
3. LITERATURE REVIEW..............................................47
The Evolution of Policy Network Approaches....................48
Major Concepts and Typology............................50
Formal Network Analysis................................54
Limitations of Traditional Policy Network Approaches...61
Policy Process Frameworks with Policy Network Perspectives....64
Advocacy Coalition Framework...........................65
Organizational State Framework.........................74
Recent Advancement of Policy Network Approaches...............84
Collective Action among Policy Actors..................84
Effects of Network Structures on Policy Outcomes.......89
Social Capital in the Policy Network Approaches........94
Macro-level Factors in the Policy Network Approaches...97
4. RESEARCH DESIGN AND METHODOLOGY...............................106
Methodological Background....................................108

Policy Networks in the Korean Foreign Labor Policy.......190
Hypotheses Test Results.........................................203
The Glue of Policy Coalitions............................203
Factors Affecting Collective Action......................212
The Effect of Trust on Mapping the Policy Networks.......217
Antecedents of Trust.....................................221
Policy Actors Influence in Policy Networks..............224
6. CONCLUSION AND DISCUSSION........................................232
Major Findings and Implications of the Research.................233
Policy Coalition in the Korean Foreign labor policy
The Glue of Policy Coalition.............................236
Collective Action........................................240
Trust in a Policy Network................................243
Policy Actors Influence.................................246
A Typology of Relationships between Policy Actors...............248
Limitations of the Study........................................251
Future Research Directions......................................253

2.1 A Research Framework for the Analysis of Korean Foreign Labor Policy
2.2 The Structure of Korean Foreign Labor Policy Network (Ko &Lee)....42
3.1 Diagram of the Advocacy Coalition Framework (2005 version).........68
3.2 Causal Model of Relations among Organizational Characteristics,
Exchange Networks, and Reputations and Event Activities...........79
4.1 Structure of Research Design......................................108
4.2 Multiple Layers around Policy Actors..............................115
4.3 Levels of Policy Network Analysis.................................118
4.4 Policy Actors Attributes and Their Effects on the Social Outcomes in a
Policy Network....................................................126
5.1 Structure of Communication Network in the Korean Foreign Labor
5.2 Structure of Ally Network in the Korean Foreign Labor Policy......197
6.1 A Model of Operating Mechanism of the Ally Network and the
Communication Network.............................................251

2.1 Changes in the Types of Foreign Workers Visa Status in Korea.....19
2.2 Trends of Labor Shortage in Korea (1985-1992)...................21
2.3 Comparison of the GNI between Korea and Labor-exporting Countries (Yr.
2.4 The Number of Foreign Workers (1990-2002).......................24
2.5 Major Changes in the ITSF.......................................29
2.6 The Structure of Korean Foreign Labor Policy Network (Seol).....44
2.7 Comparison of the ITSF and the EPSF.............................46
3.1 Types of Policy Networks: The Rhodes Model......................53
3.2 Components of the Policy Domain Framework.......................77
3.3 Coalition Behavior as a Result of Interdependency and Belief
3.4 Comparison of the ACF and the OSF.................................104
4.1 Major Elements in the Levels of Analysis.......................119
4.2 Hypotheses Regarding the Glue of Policy Coalitions.............140
4.3 Hypotheses Regarding Collective Action in a Policy Coalition...143
4.4 Hypotheses Regarding Trust in a Policy Network.................145
4.5 Hypotheses Regarding Policy Actors Influence in a Policy Network.148
4.6 Major Policy Actors in the Korean Foreign Labor Policy.........154
4.7 Linkage Information in a Network with Nonrespondents...........165

5.1 Questionnaire Response Results....................................173
5.1 Questionnaire Response Results (cond)............................174
5.2 Basic Information about the Respondents (Questionnaire Q. 33 38).... 176
5.3 Policy Actors Motives for Participation in the Policy Process
(Questionnaire Q. 22)............................................178
5.4 Policy Actors Entry Stage in the Policy Process (Questionnaire Q. 22).179
5.5 Policy Actors Advocating Interests in the Policy Process (Multiple-
responses) (Questionnaire Q. 24)..................................180
5.6 Policy Actors Major Roles in the Policy Process (Multiple-responses)
(Questionnaire Q. 25).............................................180
5.7 Policy Actors Affiliation Types and Formal Voting Power.........183
5.8 Policy Actors Trustworthiness and Influence.....................186
5.9 Summary of Collective Action Types (Multiple-responses)...........187
5.10 Policy Actors Collective Action Types...........................189
5.11 Policy Actors Network Attributes in the Egocentric Communication
5.12 Policy Actors Network Attributes in the Egocentric Ally Network.199
5.13 Comparison Major Factors between the Communication Network and the
Ally Network......................................................202
5.14 The Glue of Policy Coalitions in the Ally Network (N=174).........204
5.15 The Glue of Policy Coalitions in the Communication Network
5.16 The Effects of Each Level of the PBS on Relationship between Policy
Actors (N=174)....................................................210

5.17 Factors Affecting Communication Density between Policy Actors
5.18 Factors Affecting Relationship between Policy Actors............218
5.19 The Effect of Trust on Mapping the Ally Network..................220
5.20 The Analysis of the Factors Affecting Trust between Policy Actors
5.21 Pearson Correlations between Policy Actors Institutional Attributes and
Influence (N=25)..................................................225
5.22 Pearson Correlations between Policy Actors Network Attributes and
Influence in the Communication Network (N=25).....................227
5.23 Pearson Correlations between Policy Actors Network Attributes and
Influence in the Ally Network (N=25)............................ 227
5.24 Summary of Hypotheses Test Results...........................229
5.24 Summary of Hypotheses Test Results (cond).......................230
5.24 Summary of Hypotheses Test Results (cond).......................231
6.1 Typology of Relationships between Coalition Members................249

Advocacy Coalition Framework
Actor-Centered Institutionalism Framework
Prof. Dong-Hun Seol
Employment Permit System for Foreigners
Federation of Korean Industries
Federation of Korean Trade Unions
Grand National Party
Industrial Training System for Foreigners
Joint Committee of Industrial Training Companies for Foreigners
Korean Confederation of Trade Unions
Korean Federation of Small Businesses
Korean Joint Committee for Migrant Workers
Korean Joint Committee for measures against forced expulsion
of foreign workers, of abolishing industrial foreign training system
and of foreign workers human rights
Korea Labor Institute
Korean Office of the President
Korea Small Business Institute
Ministry of Commerce, Industry and Energy
Ministry of Finance and Economy

MOJ Ministry of Justice
MOL Ministry of Labor
NACIEC National Assembly Commerce, Industry and Energy Committee
NAELC National Assembly Environment and Labor Committee
NHRC National Human Rights Commission
NMDP New Millennium Democratic Party
OGPC Office for Government Policy Coordination
OSF Organization State Framework
SDK Prof. Soo-Dol Kang
SMBA Small and Medium Business Administration
SNA Social Network Analysis
YBL Prof. Yun-Bo Lee
YBP Prof. Young-Bum Park

Knowledge in the International Migration Policy Process
Along with the advancement of globalization, the issue of international
migration has come to occupy an extraordinary place all around the world (Bauer et
al., 2000; Klingner & Hugill, 2008; SOPEMI, 1999; Zolberg, 2006). Many
developed countries could not help importing foreign workers to maintain their
economies. Conversely, a great number of developing countries have shown more
and more interest in exporting their workers. Such a phenomenon is well illustrated
by the statistic that the number of persons who live outside their country of birth
doubled in number during the years between 1975 and 2000 (Meyers, 2004).
Moreover, recent trends in international migration indicate that the increase of
international migration will continue for the time being: i.e., more and more
workers are expected to leave their own countries to get better jobs in foreign
countries. Accordingly, it carries great weight to examine the contents and
processes of migration policies. In particular, considering that this issue is greatly
influenced by the historical background and cultural traits of each country, policy

cases of various countries are certainly expected to bring about better migration
policies as well as better understanding of the issues inherent in international
migration (Klingner & Hugill, 2008).
International migration is not a new topic for many countriesespecially,
for those countries constructed by settlers, such as the United States, Canada, and
Australia. Therefore, the studies on the evolution of international migration policies
of individual receiving countries have been abundant up to now (Tichenor, 2002;
Zincone & Gregorio, 2002; Zolberg, 2006). As some scholars (Meyers, 2004;
Skeldon, 2007) have already observed, however, the arena of international
migration policy lacks not only a systematic approach based on a general theory
but also comparative analyses across countries in a similar situation. To overcome
these limitations, some scholars (d Almeida, 2009; Meyers, 2004; Zolberg, 2006)
have tried to suggest a theoretical framework representing the factors contributing
to the formulation and change in international migration policy. For example,
Meyers (2004) attends to the domestic socioeconomic and foreign policy factors
and the type of immigration the country takes, arguing that international migration
policy is largely the result of interaction between socioeconomic and foreign policy
factors and the type of immigration, and that there are some undeniable similarities
among the international migration policies of Western countries.1
1 Meyers (2004) performed a comparative study on four Western countries: the United
States, the Britain, the Netherlands, and Germany.

Although recent developments on theorizing the international migration
policy are apparently noteworthy, the greatest weakness of the literature on
international migration policy is its lack of the consideration of the policy process,
in which policy actors make real interactions and produce ultimate policy outcomes.
Although some scholars (Skeldon, 2007; Zincone & Gregorio, 2002) have dealt
with the policy process of international migration policy, these studies focused
more on describing the process with empirical data than on testing a theory or
formulating a new one. In other words, recent studies on international migration
policy do not contribute to illuminating the black box in the policy systems with
systemic insight (Easton, 1965).
Knowledge of the International Migration Policy Process
As Lasswell (1970) emphasized, we need knowledge of the policy process
as well as knowledge in the process to get a complete understanding of the public
policy (deLeon, 1999). This dissertation tries to add values to the extant
international migration policy studies by focusing on the process of the Korean
foreign labor policy; in particular, this research pays special attention to the policy
actors and their relations in the process of adopting the Employment Permit System
for Foreigners (EPSF) during the early 2000s.
In examining the process of Korean foreign labor policy, this dissertation
attends to the policy networks in it. Specifically, it focuses on the ally network as

well as the communication network in the policy process. Accordingly, the
relationship between policy actors and collective action in the policy networks are
major research targets. Also, in parallel with the implications from the concept of
social capital, the roles of trust between policy actors in the policy process are
As Hanf (1978) already observed, some 30 years ago, public policy
processes could be characterized by the increasing involvement of different
governmental agencies and by growing interactions among government agencies,
non-governmental organizations, firms and so on. Also, considering the
implications from the concept of hollow states, in which few policy entities have
the power and authority to achieve their goals on their own, further studies on the
networks among the diverse policy participants certainly promise to give much
insight into the advancement of the policy process theories (Hanf, 1978; Milward &
Provan, 2000). In this sense, many scholars have dealt with the topic of policy
networks up to now, accomplishing meaningful advances in this area (Freeman,
1965; Heclo, 1978; Jordan, 1990; Milward & Provan, 2000; OToole, 1997;
Rhodes; 1990). They emphasize in common the significance of relationships
among policy actors, thus trying to overcome the limitations of policy studies that
solely rely on behavioral or institutional perspectives (Howlett, 2002). Also, policy
network approaches imply the necessity of incorporating network perspectives into
the policy process frameworks. For example, Dowding (1995) suggested the

following two aspects as key characteristics of the policy network approaches: (1)
the pattern of relationships among policy actors greatly influences policy outcomes;
and (2) the sub-governmental level is a significant unit of analysis, as it reveals the
details of the policy process.
However, despite the potential of the policy network approaches, some
issues should be much more developed to enhance the utility of the policy network
approaches. For example, critics (Blom-Hansen, 1997; Dowding, 1995; Klijn &
Koppenjan, 2000; Marin, 1990; Peters, 1998) point out if policy networks are to be
effective predictors of policy outcomes, the features of networksnot just policy
actors characteristics in the networksshould have strong explanatory power. To
contribute to overcoming the limitations of the policy network approaches, this
research pays particular attention to the concepts of policy networks and policy
coalitions in policy processes and attempts to develop the concepts by examining
the factors influencing the relationship and collective action among the policy
This dissertation starts with the assumption that a better appreciation of the
concept of policy networks and policy coalitions can greatly contribute to
developing theories on policy processes (Sabatier, 1993; Innes & Booher, 2003;
Milward & Provan, 2000). Along with the advancement of policy network
approaches, this dissertation endeavors to expand the applicability of policy
networks by deepening our understanding of the factors influencing the networks.

In particular, paying special attention to the limitations of the Advocacy Coalition
Framework (ACF), this thesis examines the potential of the concept of policy
coalitions as a promising unit of analysis in studying policy processes (Sabatier &
Weible, 2007). Considering that the whole policy network is often too big a unit to
handle and that individual policy actors as a unit of analysis have limits in showing
the dynamics of interaction among the policy actors, policy coalitions can be a
good unit of analysis in uncovering the mechanism embedded in a policy network.
Specifically, this dissertation develops diverse issues around policy coalitions, such
as the factors influencing the relationship among coalition members, collective
action among policy actors, and power relations in the coalition. By increasing our
knowledge about policy actors behaviors in a policy network, this dissertation will
surely contribute to producing better general policy process theories, as well as
elucidating the policymaking process of Korean foreign labor policy as an example
of international migration policy.
Research Purposes and Questions
Public policy-making processes are very complex phenomena, in which
many policy actors with different goals and resources are involved to realize their
own interests, but also diverse policy environmental factorssuch as political
structure, socio-economic situation, and national cultureaffect the processes as
well. These characteristics make it difficult to grasp the whole picture of policy

processes and to predict policy outcomes. Nevertheless, a variety of policy process
frameworks and theories have been devised and developed to explain policy
phenomena up to now; some of them (Brewer & deLeon, 1983) focus on the
functions in the public policy processes, others (Jordan, 1990; Knoke, 1990; Marsh
& Rhodes, 1992) address the policy networks in the processes, others (Ostrom,
1999; Scharpf, 1997) focus on the institutional arrangement in the policy process,
and still others (Easton, 1965) on the systems around the policy process. These
efforts have contributed to enhancing our understanding of the policy world.
However, much more development should be attained to achive better explanation
and prediction of policy outcomes.
Reflecting on extant literature in this arena, this dissertation develops the
concepts of policy networks and policy coalitions in the policy-making process.
More specifically, it considers the following four main purposes. First, this
dissertation aims to expand the understanding of a specific issue-area, i.e.,
international migration policy. Studies on the contents of international migration
policy are abundant, but those on the processes of international migration policy
are rare. The Korean foreign labor policy case of this dissertation, therefore,
substantially contributes to broadening our perspectives in terms of the policy-
making process of international migration policy. In particular, as this study adopts
analytical methods based on social network analysis, it can describe the whole

policy process with relational data, and this attempt seems to be the first trial in the
international migration policy arena.
Second, focusing on policy networks in the policy-making process, this
dissertation tries to recognize the mechanism underpinning the interactions among
policy actors in the networks. More specifically, paying special attention to the
diversity of policy network types, the thesis focuses on two specific types of policy
networks: the ally network and the communication network. These networks are
constructed upon such social outcomes as relationship and communication density
between policy actors. As a first stage of mapping the policy networks, this thesis
explores the glues of policy coalitions in these networks and identifies significant
factors for determining collective action between policy actors.
Third, the significance of trust is well recognized in the present literature of
policy studies. Thus, it is also anticipated that trust play a powerful role in
determining relationships and collective action among policy actors. This thesis
examines the roles of trust in the process of producing social outcomes in the
networks. Along with such expectation, this dissertation also attempts to find the
antecedents of trust in the policy network. Considering the rareness of this kind of
study based on network analysis, this research will surely contribute to deepening
our understanding of trust, especially in the context of policy-making process
(Robertson et al., 2007).

Finally, to obtain a more complete understanding of policy network, its
power relationships should be taken into account. In this sense, this dissertation
pays particular attention to the factors affecting policy actors influence in a policy
network. In particular, policy actors institutionalattributes, such as trustworthiness
and affiliation types, are compared with their network attributessuch as centrality,
betweenness, and closenessin terms of their comparative utilities in predicting
their influence level in the network.
Parallel with these research purposes, it should be noted that this thesis is
unique for several other reasons. Most of all, considering that major development
of policy process theories has been accomplished in the context of Western
countries, it is worthwhile to examine their utility in a different configuration or a
nation-state. For example, whereas the dominance of governance is considered a
general trend in the U.S. and the European countries, its relevance to Korea is
controversial because of Koreas long tradition of hierarchical culture. Therefore,
this study is expected to broaden the applicability of policy network approaches
into different institutional settings.
Also, the Korean foreign labor policy includes a wide range of policy
actors, such as political parties, legislative committees, various governmental
agencies, interest groups, business associations, labor unions and academics. This
kind of strength of the Korean foreign labor policy case is expected to illustrate
diverse interactions among policy actors and coalitions, which is essential to the

study of policy networks. In addition, as Korea has shown a rapid development not
only in the economic area but also in the political area since the early 1970s, it
should be noted that the status of Korea in terms of international migration policy
has been greatly changed since then; i.e., Korea was one of the major labor
exporting countries during the 1970s but from the latter part of the 1980s, Korea
has dramatically transformed into a foreign labor importing country due to its rapid
economic growth. Thus, the foreign labor policy process of Korea is expected to
illustrate dynamic changes in both of the contents and processes of the international
migration policy.
The following four major research questions present the objectives and
directions of this dissertation in depth:
(1) What are the glues of policy coalitions in policy networks? And is the
concept of policy actors belief systems effective in determining policy
(2) What are the influential factors for determining the degrees of collective
action between policy actors?
(3) What are the roles of trust between policy actors in shaping the policy
networks? Do the variables of the proximity of belief systems, the
proximity of policy interests, and the proximity of policy preferences
between policy actors influence the communication density and the
relationship between policy actors when the levels of trust between

policy actors are controlled? Also, what are the antecedents to trust in a
policy network? In other words, what factors are significant for
determining trusting relationships between policy actors in a policy
(4) How can we distinguish more influential policy actors from less
influential ones in the policy process? Which one is more influential for
determining the levels of policy actors influence in a policy network
between policy actors institutionalattributes and their network
Structure of the Dissertation
This dissertation consists of six chapters including this introduction. The
second chapter introduces the evolution of Korean foreign labor policy, focusing on
Koreas legal systems, policy environments, and major policy changes. This
chapter also shows the major issues of the Korean foreign labor policy and depicts
the evolution of the policy, especially focusing on the policy alternatives and policy
actors activities during the policy process in 2002 and 2003.
The third chapter presents a literature review, paying special attention to the
literature on the policy network approaches. The evolution of policy network
approaches is reviewed along with the limitations and recent progress of the
approaches. In the chapter, diverse arguments regarding policy coalitions and

policy networks are focused upon to appreciate the factors influencing the
relationship and collective action among policy actors in the networks.
The fourth chapter presents the research design and methodology. In this
chapter, the theoretical framework is first presented on the basis of the relevant
literature review along with major variables for this research. Research hypotheses
are then put forward with brief explanations of the sources of the hypotheses. An
explanation about data collection and data analysis follows the discussion of the
validity and reliability of this study. Data collection and data analysis sections show
that this thesis depends chiefly on the data collected by questionnaires, and that the
data are analyzed with the UCINET and SPSS programs.
The fifth chapter presents the analysis results of this research. In the first
part, descriptive analysis results are shown, focusing on describing major attributes
of the policy actors in the policy process and on mapping two major types of policy
networksthe ally network and the communication network. For this, the UCINET,
a network-based software, is adopted to analyze the relationship among the policy
actors. The next part reports the test results of the proposed hypotheses. This part is
based on quantitative analyses such as multiple regression and correlation analysis.
The hypotheses about the relationship among major variables are examined using
the SPSS software.
Lastly, the sixth chapter summarizes major findings of this research and
suggests the implications of the findings. Also, the significance of this research is

underlined again along with the limitations of the study and the directions for future

The beginning of Korean foreign labor policy traces back to the late 1980s.
With its constant economic growth since the 1970s, the Korean economy did not
have the manpower necessary to avoid severe shortage of unskilled workers at that
time. Since then, the issue of importing unskilled foreign labor has produced a
major controversy in Korea (Ha & Choi, 2005; Uh, 1999; Y. Kim, 1994).
Those who favored the influx of unskilled labor argued that foreign
workers would not reduce the job opportunities for the domestic workers because
Korean workers would not take those jobs filled by foreign workers due to low
wage and poor working conditions (Ha & Choi, 2005; Seol, 1999). They also
argued that the employment of unskilled foreign workers was indispensable for
both the Korean companies and their domestic workers to survive in a competitive
2 In general, the targets of foreign labor policy include both professional foreign labor and
unskilled foreign labor. This dissertation focuses exclusively on unskilled foreign labor
policy in Korea.

world economy because many of the small and medium businesses could not afford
to relocate their factories overseas to access lower-wage workers.
On the contrary, those who opposed the immigration of foreign workers
worried about the side effects of the trend (Ha & Choi, 2005; Uh, 1999). They
maintained that those companies depending mainly on cheap labor would not
survive in the world economy, eventually weakening the basis of the Korean
economy. This argument was supported on the grounds that the Korean economy
should make a transition from its existing low-wage and low-tech industries to new
industries with high-tech and high value-added products. The utilization of the aged
population and women, they suggested as an alternative, were two under-employed
segments of the population that could reduce the labor shortage (Uh, 1999).
Based on these arguments, Korean foreign labor policy has been changed
from the late 1980s. Roughly, the evolution of foreign labor policy in Korea can be
summarized as a competition between two policy programs: the Industrial Training
System for Foreigners (ITSF) and the Employment Permit System for Foreigners
(EPSF). The ITSF was the first policy program adopted in Korea to satisfy growing
domestic demands for unskilled foreign workers, but as the ITSF was nominally a
training programi.e., not a program for the official immigration of foreign
workersit caused many problems. For example, the Labor Law did not protect
foreign industrial trainees adopted under the ITSF because they were not workers
in a legal sense. Also, the ITSF was criticized for entailing the rapid increase of the

unskilled foreign workers illegal stay. The problems of the ITSF facilitated a
search for policy alternatives, and the EPSF came onto the stage in the mid-1990s
as the most prominent alternative to the ITSF. However, many small business
owners and their associations strongly opposed the adoption of the EPSF, on the
grounds that it would increase the cost of hiring foreign workers.3 As a resu It, the
first trial of the enactment of the EPSF between 1995 and 1997 failed. Instead, a
new programWorking after Training Program for Foreigners, which was a
revision of the ITSFwas adopted as a second best option. Another round of
policy debate was started in 2000, and after fierce debates on the issues around the
adopting of the EPSF policy, the Korean National Assembly finally decided to
adopt the EPSF system in July 2003, on the condition that it would coexist with the
ITSF for the time being. This chapter tries to explain the changes in the Korean
foreign labor policy more concretely with an analytical framework, which will be
introduced in the next section. In particular, this chapter focuses upon the policy
process during 2002 and 2003, which is the research target of this thesis.
3 It may sound strange that small business owners opposed the adoption of the EPSF. In
general, business groups assent to expand the import of foreign manpower because it
decreases labor cost. But it should be noted that the ITSF was more useful for the small
business owners than the EPSF in terms of labor cost reduction at that time because the
ITSF did not provide foreign workforces with legal labor rights, such as minimum wage
and labor standards (e.g., working hours, holidays, etc.)

Analytical Framework
The evolution of the Korean foreign labor policy is analyzed with the
following framework, developed from the Advocacy Coalition Framework (ACF)
(Sabatier & Weible, 2007). Figure 2.1 shows one conceptual scheme of the
Source: modified from Sabatier and Weible, 2007, p. 202.
Figure 2.1. A Research Framework for the Analysis of Korean Foreign Labor
Policy Process.
In the framework, first, two kinds of environmental factors are identified as
the socio-economic environment and political environment. Socio-economic
environment includes overall economic conditionssuch as GDP growth, labor

shortage and unemployment rateand the size of the unskilled foreign worker pool.
Political environment pays special attention to the changes in the Parliament and
the Executive government. Another significant element in this framework is
focusing events. It is not rare that unexpected focusing events dramatically change
the fundamental grounds of a policy system (Kingdon, 2003). In this sense, critical
events during the policy process are identified and analyzed. The characteristics of
policy problems, which are greatly affected by the changes of exogenous policy
environments and focusing events, are then analyzed. In analyzing this element, the
effects of previous policies on defining the policy problems are fully considered.
The changes in the policy problems reconstruct the policy subsystem in many ways.
Sometimes, for example, new policy actors enter the process and change the power
structure among policy coalitions in the policy subsystem. Finally, the changes in
the power structure among policy coalitions bring new policy outcomes and policy
effects. One important result related to policy effects is that new policy outcomes
generate renewed mechanisms of gains and losses, which also affects
organizational interests.
Basic Legal Systems Regulating Foreign Workers
In Korea, the Ministry of Justice has the jurisdiction over immigration
control by enforcing the Immigration Control Act. Foreign workers can be grouped
into three categories by the standard of their visa status: the legally employed; the

industrial trainees; and the illegal foreign workers. As Table 2.1 shows, there was
no official visa status for unskilled foreign workers in Korea before 1991.
Table 2.1
Changes in the Types of Foreign Workers Visa Status in Korea
Legal Foreign Workers (E type Visa)
Year Professional and Technical Workers (El E7)4 5 Production Workers after Training (E8)6 Non- professional Workers (E9)7 Industrial Foreign Trainees (D -3/
Before 1991 V - - -
1991 1997 V - - V
1998-2002 V V - V
2003 - V V V V
Source: Ha & Choi, 2005.
4 As the Korean government had prohibited the influx of unskilled workers until the
adoption of the E-8 visa in 1998, the only legal way for unskilled foreign workers to work
in Korea was to acquire this type of visa status.
5 Th e group of professional and technical workers includes university professors (E-l),
language teachers (E-2), researchers (E-3), technology instructors (E-4), professionals (E-
5), entertainers (E-6), and those under specific activities (E-7) who cannot be replaced by
Korean workers belong to this visa status.
6 Th e visa for production workers after industrial training (E-8) was adopted to
supplement the Industrial Training System for Foreigners (ITSF) in 1998. The adoption of
E-8 visa meant that the Korean government institutionalized the import of unskilled
foreign workers for the first time.
7 Th e passage of the Employment Permit System for Foreigners (EPSF) in 2003 brought
the new visa status of E-9 for those who entered Korea to obtain non-professional jobs
under the regulation of the Act on the Foreign Workers Employment, etc.

The adoption of the ITSF in 1991 created a new type of visa statusD-3 visa for
Industrial foreign trainees. Then, Working after Training Program for Foreigners,
adopted in 1998, brought the first official visa status (E-8) for unskilled foreign
workers. Lastly, the adoption of the EPSF brought the E-9 visa status in 2003.
Major Changes in the Korean Foreign Labor Policy8
As we know from the changes in the types of foreign workers visa status,
there were three major changes in the Korean foreign labor policy since the latter
part of 1980s. The first round (1989 1991) yielded the ITSF. The second round
(1994 -1998) concluded in the adoption of the Working after Training Program for
Foreigners. Finally, the third round (2000 2003) legalized the direct import of
unskilled workers with the enactment of the EPSF. The following section gives a
full account of these three policy rounds in turn.
The First Round (1989-1991): Adoption of the Industrial Training System for
The origin of Korean foreign labor policy can be traced back to the late
1980s. The migration of unskilled foreign workers was not a serious social agenda
until that time. However, as Korea has developed into an industrialized economy
8 In this paper, foreign workers refer to those foreigners who provide their labor for wage,
regardless of their legal visa status. Thus, the category of foreign workers includes
industrial foreign trainees and illegal foreign workers, as well as legal foreign workers.

with relatively high wages and relatively good working conditions compared to
other developing countries,9 the Korean economy has faced a severe shortage of
unskilled workers in its small and medium businesses across industries, especially
in the so-called 3D (Dirty, Dangerous, and Difficult) occupations.10 T able 2.2
illustrates the seriousness of labor shortage in Korea during the latter part of the
1980s and early 1990s. Note that the labor shortage in 1991 became about five
times larger than that of 1985, threatening sustainable growth of the Korean
Table 2.2
Trends of Labor Shortage in Korea (1985-1992)
Unit: person
Total Number of Labor Shortage Professional Technician and Similar Occupations Production and Similar Occupation
1985 54,706 11,225 43,481
1987 121,026 17,722 103,804
1989 141,419 21,572 119,847
1990 192,055 26,203 165,952
1991 250,108 28,323 221,785
1992 196,563 39,668 156,895
Source: adapted from Kim, Y., 1994.
9 By 1995, according to World Bank figures, the size of the Korean economy was ranked
12th in the world (World Bank, 1995).
10 Th e 3Ds occupations typically lie in the areas of dying, plating, heat-treating, casting
and tempering, footwear, glass, and leather.

In particular, the 1986 Asian Games and the 1988 Olympic Games hosted
by Korea played an important role in advertising the striking economic
accomplishment of Korea and attracted a great number of foreign workers to Korea.
In addition to the economic attractions (i.e., relatively high wages and abundance of
job opportunities) of Korea, the changes in social and international environments
also prompted the growing influx of foreign workers. For example, the Korean
government revised the Immigration Control Law in 1992 to help foreigners visit
Korea more readily. Also, the changes in international environments, such as the
outbreak of the first Gulf war in 1990 and the Japanese governments strict
regulation on foreign workers influx to Japan in 1989, narrowed down the target
countries in which the international migrant workers might choose to work, making
Korea particularly attractive as a new migrant country (Seol, 1999).
Despite relatively poor working conditions in the jobs in which there were
labor shortages, foreign workers increasingly emmigrated to Korea and took these
jobs willingly because their wages were very high compared with those in their
home countries. As Table 2.3 indicates, the gap of gross national income (GNI)
between Korea and labor-exporting countries was significantly large. For example,
the GNI of Korea was more than 10 times that of the Peoples Republic of China in

Table 2.3
Comparison of the GNI between Korea and Labor-exporting Countries (Yr. 2000)
Country GNI (US $) Proportion to the GNI of Korea (%)
Republic of Korea 8,910 100
China 840 9.4
Thailand 2,010 22.6
Philippines 1,040 11.7
Pakistan 470 5.3
Sri Lanka 870 9.8
Bangladesh 380 4.3
Source: World Bank, 2001, pp. 1-3.
As a consequence, Korea became an increasingly attractive country for
unskilled foreign workers, who began to occupy a large portion of the Korean
domestic labor market. Also, as the labor shortage of unskilled workers became
severer and the wages for domestic workers went higher, Korean employers
increasingly demanded many more unskilled foreign workers whose average wages
were much lower than those of Korean workers. Table 2.4 illustrates the rapid
increase of unskilled foreign workers since 1990. The total number of foreign
workers in 2002 became almost 18 times larger than that of 1990.

Table 2.4
The Number of Foreign Workers (1990-2002)
Unit: person
Year Legal Foreign Workers Industrial Trainee Undocumented Foreign Workers Total
1990 2,833 0 18,402 21,235
1991 2,973 599 41,877 45,449
1992 3,395 4,945 65,528 73,868
1993 3,767 8,048 54,508 66,323
1994 5,265 24,050 48,231 77,546
1995 8,228 52,311 81,866 142,405
1996 13,420 68,020 129,054 210,494
1997 15,900 81,451 148,048 245,399
1998 11,143 47,009 99,537 157,689
1999 12,592 69,454 135,338 217,384
2000 19,063 77,448 188,995 285,506
2001 27,614 46,735 255,206 329,555
2002 33,697 39,661 289,239 362,597
Source: Ministry of Justice, 2003
For a better appreciation of the changes in the Korean foreign labor policy,
a more thorough explanation of the illegal foreign workers is necessary because
their proportion of the total foreign workers has increased disproportionately since
1990. Illegal foreign workers had typically come to Korea because of increased
employment opportunities since the late 1980s when wages for domestic workers

increased and the labor shortage became much more crucial for the Korean
economy. The rapid increase of illegal foreign workers was encouraged by the lax
law enforcement of the Korean government. Korean government authorities tacitly
permitted illegal foreign workers as a means for balancing the labor market, either
by passing over their illegal residency or by strengthening law enforcement,
depending on the economic situation of Korea. For example, the number of illegal
foreign workers decreased from 148,048 in 1997 to 99,537 in 1998. At that time,
the Korean economy suffered from a serious economic crisis and as a result, the
unemployment rate soared up to 7.0% in 1998. Responding to the economic crisis,
the Korean government devised a policy to assist those employers who would
replace illegal foreign workers with Korean workers at that time and strengthened
law enforcement on illegal foreign workers.11
The number of undocumented foreign workers based on the statistics of the
foreigners departure and arrival records indicates the approximate figures of illegal
foreign workers. Table 2.4 shows the number of illegal foreign workers from these
statistics. According to the Ministry of Justice (2003), most of the workers came
from the developing countries such as the Peoples Republic of China, the
Philippines, Bangladesh, Thailand, Mongolia, Vietnam, and Indonesia.
11 Th e unemployment rate of Korea was 2.6% in 1997 but it became 7.0% in 1998, then
6.3% in 1999 and 4.1% in 2000. Since 2000, the rate has been stabilized at the level of 3%.

In terms of political environment, the year 1987 marked a new epoch in
modem Korean politics. After a long rein of military regimes since the early 1960s,
in 1997, a strong peoples movement for democracy changed the Korean
constitution and enabled the direct election of a President by the Korean people.
Although the result of the presidential election in 1987 did not bring a change of
regime, the political environments was much more improved towards a basic
political democracy. As a result, the influences of labor unions and non-
governmental organizations were strengthened. For example, the number of labor
unions strikes dramatically increased during the period 1987-1989, mainly due to
the changing political environments: the number was only 276 in 1986, but it
soared to 3,749 in 1987, then 1,873 in 1988 and 1,616 in 1989. Such changes in the
political environment made it difficult for the Korean government to make a policy
decision without obtaining major stakeholders support.
Under these economic and political environments, the policy issue of
importing unskilled foreign workers came into the policy stage at the request of
some business interest groups. For instance, the Korean Association of Coal
requested the importation of foreign workers to the Ministry of Energy and
Resources in October 1989, and the Cooperative Association of the Electronics
Industry made the same request in 1990. Their requests aimed to solve the problem
of labor shortage and were supported by the Ministry of Commerce and Industry.

When the issue of importing unskilled foreign workers first entered the
policy stage in the latter part of the 1980s, many related policy actors (including the
Ministry of Labor, the Ministry of Justice, the Economic Planning Board and major
labor unions) were staunchly opposed to importing foreign workers because they
worried about the side effects foreign workers would entail. For example, the new
supply of foreign workers would decrease job opportunities for domestic workers
and increase the social costs for managing foreign workers (Y. Kim, 1994). At the
same time, however, Korean policy makers could not ignore domestic employers
demand for unskilled foreign workers, which were represented by the Ministry of
Commerce and Industry and the Federation of Korean Industries.
Responding to the major policy actors opposition to the importation of
foreign workers, the Ministry of Commerce and Industry found a solution in the
Japanese model of the Industrial and Technical Training Program for Foreigners
(Park, 1992). Formally, the adoption of the ITSF was not intended to legalize the
importing of foreign workers because its official goal lay more in the occupational
training of foreigners than in utilizing foreigners as members of the workforce.
However, almost all the policy actors were also aware of the possibility that foreign
industrial trainees would actually work for the training companies, following the
Japanese model (Seol, 1999). That is, the trainees were expected to contribute to
reducing labor shortages by performing work in their respective factories.

As a result, the ITSF was adopted as a political solution for reducing the
labor shortage at that time. It was a primarily political solution, in the sense that the
ITSF could at least partly satisfy major policy actors. The opponents of importing
foreign workers could justify their support for the agreement because the ITSF was,
at least formally, not a program to institutionalize the importation of foreign
workers. The proponents expected to earn real benefits by adopting the ITSF
because they knew that the ITSF would substantially contribute to mitigating the
Korean labor shortage. Such mutually appealing traits of the ITSF, on the one hand,
helped to serve as a compromise among related policy actors at that time. On the
other hand, these same characteristics became a departure point between the
different factions during the long controversy over proper legal systems regulating
unskilled foreign workers.
The ITSF started with a small coverage population, targeting only those
foreign workers who were employed by Korean companies established outside
Korea. Only those companies were allowed to bring their foreign trainees into
Korea. Also, the number of foreign trainees was confined to less than 50 persons or
10 percent of the number of Korean workers in the same company. The training
could last six months with a possible extension of an additional six months with
explicit consent from the Ministry of Justice (Ha & Choi, 2005). As the labor
shortage of unskilled workers became more severe, however, the ITSF expanded its
coverage again and again by executive orders. As a result, the number of industrial

trainee jumped from 599 in 1991 to 81,451 in 1997. Table 2.5 shows a summary of
the major changes in the ITSF.
Table 2.5
Major Changes in the ITSF
ITSF under the Government ITSF under the KITCO
Is* 2nd 1st 2nd 3rd 4 5th
Round Round Round Round Round Round Round
Policy Oct. Aug. Nov. Sep. May Feb. June
Making 1991 1992 1993 1994 1995 1996 1996
Enforce- Nov. Sep. May Dec. Sep. June Dec.
ment 1991 1992 1994 1994 1995 1997 1996
Planned Number Up to 10% of Perma- 10,000 20,000 Garment: 7,500 20,000 19,000 10,000
of Trainees nent Workers Footwear: 2,500
Training Period 6 months 6 months 1 year 1 year 1 year 2 years 2 years
Sources: Adapted from Seol, 1999, p. 434.
Note. ITSF means Industrial Training System for Foreigners and KITCO is the
abbreviation of Korea International Training Cooperation Corps.
The Second Round (199 -1998): Emergence of the EPSF as an alternative to the
The start of a new policy round was initiated by the negative effects of the
ITSF and some unexpected focusing events due to these issues. Meanwhile, the
economic and political environments were essentially similar to the previous policy
round. First of all, the adoption of the ITSF brought about some significant

negative effects, although the policy contributed more or less to reducing the
countrys labor shortage. One of the biggest problems came from the rapid increase
of illegal foreign workers despite the adoption of the ITSF. Although Korea had
become a popular destination for international migrant workers since 1990, the
Korean government did not provide sufficient legal opportunities for foreigners
who wanted to work in Korea. The only exception was the ITSF, but despite that
policy, it was still more attractive for foreign workers to stay illegally in many
respects. Two of the most important reasons for the increase of illegal foreign
workers were the following: (1) the wages for illegal foreign workers were higher
than those of the industrial foreign trainees and (2) the risk of being expelled by
Korean officials was low because of lax law enforcement of labor laws. As a result,
many foreign industrial trainees also wanted to become illegal foreign workers
because they could earn more money. Another problem emerged due to the absence
of a legal protection system of foreign workers labor rights. Typically, some
employers encroached upon foreign workers rights by unfairly exploiting their
illegal residency status. Some chief encroachments included delays in wage
payment, forced overtime work, and improper medical treatment after industrial
accidents (Ha & Choi, 2005; Huh, 2004).
With those side effects of the ITSF, a few focusing events occurred during
1994 and 1995 that contributed to publicizing the problems of the ITSF. In 1994,
some foreign workers were injured by industrial accidents and demonstrated to

request proper compensation and protection against such accidents. Also, 13
foreign workers at Myeongdong Cathedral in January 1995 protested against unfair
treatment at the hands of their Korean employers, such as beatings, delayed
payments, and forced overtime.12 13 As the colle ctive actions by foreigners were very
rare in Korea, these focusing events were very shocking and impressive to Koreans,
1 o
with the general public being very sympathetic toward them. The events
encouraged some religious organizations and non-governmental organizations to
establish a new umbrella organization, the Korean Joint Committee for Migrant
Workers (KJCMW), in July 1995. From then on, it played a significant role in
adopting the EPSF (Huh, 2004).
The influence of those events greatly changed the characteristics of the
policy problem. Whereas the policy problem was first perceived as an economic
issue, it became a political one by extending its coverage to the areas of social
welfare, justice, equality, and international relations. To cope with these problems,
the Korean government decided to apply relevant labor laws to the illegal foreign
workers in 1995. As a result, the laws regulating compensation from industrial
12 One of the major characteristics of this policy round is that some religious groups
participated eagerly in the policy process. Interestingly, the Italian international migration
policy-making process also experienced similar phenomena: the Catholic Church played
significant roles in developing and changing international migration policy (Zincone &
Gregorio, 2002).
13 Schne ider and Ingrams social construction of target populations framework can be
useful for understanding the changes in the perception on the foreign workers (Schneider &
Ingram, 1997).

accidents, minimum wage, and delays in wage payment were applied to foreign
workers from that point. This change can be evaluated as a small progress toward
higher protection of foreign workers labor rights. Also, with such changes in the
social environment, the EPSF entered the policy stage as an alternative to the ITSF.
In this policy round, new members came into the policy stage. NGOs (such
as KJCMW) and the newly energized labor unions actively participated in the
policy process to protect foreign workers labor rights. In particular, the Federation
of Korean Trade Unions (FKTU) was very enthusiastic to protect foreign workers
rights. These new policy participants influenced the changes in power structure
between policy coalitions. The Ministry of Labor established a project team
concerning foreign labor policy and issued a report suggesting the EPSF as a policy
alternative. A number of NGOssuch as KJCMW and the Joint Committee for
measures against forced expulsion of foreign workers, of abolishing industrial
foreign training system and of foreign workers human rights (JCM)fought to
abolish the ITSF and to adopt a new policy for the importation of unskilled foreign
workers. Also, two Act proposals regarding the adoption of the EPSF were issued
by Congresspeople Jae-Oh Lee and Yong-Suk Bang, and two petitions for the
enactment of the EPSF were submitted by Su-Hwan Kim14 and H yeong-Kyu Park,
famous religious leaders in Korea at that time (Eluh, 2004).
14 Su -Hwan Kim was the first Korean cardinal of the Roman Catholic Church.

However, the Korean Federation of Small Businesses (KFSB) and the Small
and Medium Administration (SMA) strongly opposed the new program, arguing
that the EPSF would increase the cost of employing foreign workers and that labor
relations could be unstable due to possible collective actions by foreign workers
(Seol, 1999). For instance, the KFSB published a statement of protest registering its
strong opposition to the enactment of the EPSF in October 1996. Also, the KFSB
led diverse activities to prevent the adoption of the EPSF, including massive
demonstration against the EPSF by mobilizing collective action with its members
in May 1997 (Ha & Choi, 2005; Huh, 2004).
Strong conflict between two coalitions over the adoption of the EPSF
prevented the National Assembly from enacting the EPSF during the Assembly
sessions of 1996 and 1997. Instead, the program of Working after Training for
Foreigners was adopted as an amendment to the ITSF in November 1997. It was
the first official program that allowed the migration of unskilled foreign workers.
Under its provisions, foreign industrial trainees who could pass certain skill tests
after their two-year training period could stay in Korea as legal workers for one
additional year. For this program, a new visa status was introducedthe working
after training (E-8) category. In 2002, the maximum training period was shortened
to one year, whereas the working period was extended to two years. This program
can be evaluated as a tentative compromise to mitigate the pressure from those who
favored to adopt the EPSF.

The Third Round (2000 2003): Adoption of the Employment Permit System for
Due in part to the recovery from the economic crisis that occurred in 1997,
the political environments were favorable for those who consented to the adoption
of the EPSF. For the first time in modem Korean politics, presidential power was
shifted from the ruling party to the opposition party as a result of the 1997
presidential election. Dae-Jung Kim, who had played a critical role in the
democratization of Korea, became the President of Korea in 1998. More than
anything else, it should be noted that the new ruling partythe New Millennium
Democracy Party (NMDP)favored the adoption of the EPSF. However, the
NMDP was in the minority in the National Assembly at that time, which made it
difficult to reform the foreign labor policy. In 2002, there was another presidential
election, and Mu-Hyeon Roh, a member of the ruling party, became President of
Korea. One outstanding point related to the foreign labor policy during the 2002
presidential election was that both the NMDP and the Grand National Party (GNP),
major political parties in Korea at that time, made a public promise to adopt the
EPSF (Huh, 2004; Ko & Lee, 2004).
The issue of importing foreign workers did not attract public attention
between 1998 and 2000 because Korea was suffering from a serious economic
crisis. As the crisis entailed rapid increase of unemployment, the Korean
government tried to replace the existing foreign workers with unemployed

domestic workers. However, even high unemployment rates could not resolve the
problem of labor shortage because unemployed Koreans chose unemployment
rather than having jobs in less attractive occupations (e.g., the 3Ds occupations).
Thus, the issue of importing unskilled foreign labor recurred at the beginning of
2000 with the rapid economic recovery.
The Ministry of Labor tried to adopt the EPSF with the cooperation of the
ruling party, the NMDP. However, business organizations led by the KFSB
demonstrated strong opposition to the new policy system. For example, vice-
chairpeople of five major business associations published a manifesto in opposition
to the adoption of the EPSF in July 2000. Consequently, in spite of the large efforts
of the Ministry of Labor and the ruling party, the attempt to enact the EPSF did not
bear any fruit until 2003.
However, the demands for the adoption of the new policy system for
unskilled foreign workers continued. In December 2000, A Petition for the
Enactment of Protecting Foreign Workers employment and Human Rights was
proposed by the NGO, Joint Committee for measures against forced expulsion of
foreign workers, of abolishing industrial foreign training system and offoreign
workers human rights (Ko & Lee, 2004). Later, the Federation of Korean Trade
Unions submitted A Petition for Legislation of Act on the Employment and
Protection of Foreign Workers to the National Assembly in August 2002, and the
Korean Confederation of Trade Unions also submitted A Petition for Legislation of

Act on Foreign Workers Labor Permit and Protection of their Human Rights in
October 2002. These activities were culminated by the proposal of the Act on the
Foreign Workers Employment and Management by 33 members of National
Assembly represented by the Congressperson Lee, Jae-Jung in November 2002 (Ha
& Choi, 2005).
The victory of the ruling party in the presidential election in December
2002 facilitated the adoption of a new policy system for unskilled foreign workers.
The Office for Government Policy Coordination established a taskforce team for
the improvement of the foreign labor management system in December 2002.15
The Ministry of Labor announced the next month that it would make efforts to
adopt the EPSF and enforce it from 2004. Also, the National Human Rights
Committee recommended to the Chairperson of the Korean National Assembly and
the Prime Minister in February 2003 that they should adopt the EPSF for the
purpose of improving foreign workers human rights.
With the emergence of the new government in February 2003, the issue of
importing foreign workers was again spotlighted because the new government
promised to reform the foreign labor policy during the presidential election
campaign. The Roh government was very reform-oriented and considered the
adoption of the EPSF as one of the most important national reforms on its agenda.
15 Rela ted ministries, such as the MOF, the MOL, and the MOCIE, were the members of
the taskforce team.

However, many other policy actors still opposed the EPSF. In February 2003, five
major business associations officially objected to the adoption of the EPSF; in
particular, the Korean Federation of Small Businesses (KFSB) showed strong
opposition to the EPSF by presenting A Petition for Dissenting from the Legislation
related to the Act on Foreign Workers Employment Permit and Protection of their
Human Rights to the National Assembly.
To forge a compromise between the proponents and opponents of the EPSF,
a conference for the improvement of foreign labor management system was hosted
by the NMDP in March 2003. But the forum did not forge a compromise between
the pros and cons. Meanwhile, the Chief of the Office for Government Policy
Coordination hosted a meeting with related Vice Ministers and announced the
adoption of the EPSF in March 2003. In April 2003, President Roh, in a Cabinet
meeting, also acknowledged the necessity of the EPSF (Ko & Lee, 2004).
Reacting against the governments plan, the KFSB again declared its
opposition to the adoption of the EPSF. It suggested an action plan to oppose the
adoption, including such actions as massive demonstrations and a petition to the
National Assembly. To mitigate the strong opposition of many business
associations, the NMDP posed a compromise under which the EPSF would be
applied to specific industries to assess its strengths and weaknesses before
regulating whole industries. This time, however, the compromise brought criticism
from those who had favored the EPSF. Typically, major federations of labor unions

criticized the compromise and strongly requested the government to adopt the
EPSF immediately in April 2003. On April 7, 2003, President Roh reconfirmed the
necessity of the EPSF, indicating the details of the EPSF should be decided in the
National Assembly (Huh, 2004).
To prevent a breakthrough, the opponents of the EPSF strengthened their
collective actions. On April 16, 2003, the employers of small and medium
businesses sponsored a demonstration to resist the adoption of the EPSF. As a result,
in April 2003 the NMDP suggested another compromise, one in which the EPSF
would be enforced without abolishing the ITSF. And finally the Labor Ministry and
the KFSB implied the possibility of accepting the compromise suggested by the
NMDP in June 2003. Then, the Minister of Commerce, Industry and Energy
Ministry also agreed with the compromise suggested by the NMDP. Also, five
major business associations officially announced that they agreed with the
compromise suggested by the NMDP (Ko & Lee, 2004).
On the political side, a variety of activities were performed to make a deal
among the diverse interests. On April 17, 2003, the National Assembly
Environment and Labor Committee (NAELC) hosted a public hearing about the bill
on the Foreign Workers Employment and Management by 33 members of the
National Assembly, represented by Congressperson Lee, Jae-Jung.16 Howev er, the
16 Six witnesses participated in the public hearing: Gil-Sang Yu from the Korea Labor
Institute (KLI), Yun-Bo Lee from GunKuk University, Hawe-Sung Kim from KJCMW,

public hearing only reconfirmed the big differences between two conflicting
coalitions, making it difficult for the NAELC to pass the bill. Before making a
decision on the bill, the major political parties had to decide their positions. The
ruling NMDP had a comparatively favorable position to the adoption of the EPSF,
but as it was not a majority party in the Korean Assembly at that time, it was forced
to negotiate with the opposition party, the GNP. However, the GNP was not positive
to the adoption of the EPSF, although it agreed with the necessity of the adoption of
the EPSF during the Presidential election period. In particular, the members in the
National Assembly Commerce, Industry and Energy Committee (NACIEC)
strongly opposed the adoption of the EPSF.17 Howe ver, the compromise between
the MOL and the MOCIE weakened the ground of the opposition to the EPSF.
Moreover, as the delay of the expulsion for the illegal foreign workers was due to
be terminated at the end of August 2003, many feared that a failure to enact the
EPSF would result in chaos. To avoid the political burden from the opposition to
the adoption of the EPSF, the GNP had to decide its position. For the purpose, the
executive members of the GNP decided to conduct a survey of the adoption of the
EPSF on its congresspeople, and the majority of its congress members showed
approval of the adoption. After hosing a series of internal meetings on this issue,
Guk-Myeong Lee from the KFSB, Gil-Oh Jung from the Korean Confederation of Trade
Unions (KCTU) and Jung-Tae Kim from the Federation of Korean Industries (FKI), and
they were included in this thesis survey.
17 It should be noted that the chairman of the NACIEC at that time, Sang-Kyu Park who
belonged to the GNP, was ex-chairman of the KFSB.

the GNP finally decided to endow its Congresspeople with free votes on this issue
under the recommendation of approval on July 31, 2003 (Huh, 2004).
With the changes in the political parties, the Secretaries of the NAELC
made a deal to pass the bill with the condition of the coexistence of the EPSF and
the ITSF on July 2, 2003 and the plenary session of the National Assembly passed
the bill on the EPSF on July 31, 2003. President Roh officially announced the
enactment of the Act on the Foreign Workers Employment, etc on August 16,
2003. Accordingly, employers who could not employ Korean workers due to labor
shortage could legally hire foreign workers from August 2004 (Ministry of Labor,
It is useful to review some major studies on the Korean foreign labor policy
as a way of summarizing this chapter. There have been many studies on Korean
foreign labor policy, and the studies are helpful for identifying major issues and key
players in the policymaking process (e.g., H. Kim, 1996; Ha & Choi, 2005; Huh,
2004; Seol, 1999; Uh, 1999; Y. Kim, 1994; Y. Park, 1992). Among them, Ko and
Lees study (2004) and Seols research (1999) deserve particular attention because
they explicitly had interest in mapping the policy network focusing on the
18 Th e chief events in the third round are summarized in Appendix A.

formation and transition of coalitions in the Korean foreign labor policy process
during the 1990s and early 2000s.
Figure 2.2 shows the changes in the structure of the policy network over
time that was suggested by Ko and Lee (2004). Their analysis shows clearly the
members of each coalition and the changes in the members as time went by. Also,
they suggested an overall relationship between two coalitions in the policy process.
However, as their study was not based on robust empirical data analysis, it has
limits in explaining why some actors were grouped into key players and some
others were not. Also, they did not show why there were only two coalitions, not
three or four coalitions. Moreover, their description on the relationship between
coalitions was not grounded on solid data analysis.

Ministry of Commerce, Industry and Ministry of Labor
Energy Federation of Korean Trade
Federation of Korean Industries Weak Unions
Conflict NGOs for Human rights
(Cooperation among members) < (Cooperation among members)
In favor of adopting the EPSF Against adopting the EPSF

Ministry of Labor Ministry of Trade and Industry
Federation of Korean Trade Unions Small and Medium Business
NGOs for Human rights Conflict Administration
National Assembly Environment and < Ministry of Justice
Labor Committee Korean Federation of Small
(Cooperation among members) > Businesses (Cooperation among members)
In favor of adopting the EPSF
Against adopting the EPSF

Ministry of Labor Ministry of Trade and Industry
Federation of Korean Trade Unions Small and Medium Business
NGOs for Human rights Conflict Administration
National Assembly Environment and < Korean Federation of Small
Labor Committee Businesses
New Millennium Democratic Party (Cooperation among members) > (Cooperation among members)
In favor of adopting the EPSF
Against adopting the EPSF
Source: Modified from Ko & Lee, 2004, pp. 35-36.
Figure 2.2. The Structure of Korean Foreign Labor Policy Network (Ko & Lee).

Another significant effort to map the Korean foreign labor policy network
was undertaken by Seol (1999). He mapped the policy network around the issue of
importing foreign workers during the 1990s. In the study, Seol categorized two
rounds of policy debate and identified major policy actors. Table 2.6 shows the core
of his study.

Table 2.6
The Structure of Korean Foreign Labor Policy Network (Seol)
Policy Issue Import of Foreign Workers Enactment of Special Law on Foreign Workers
Period 1990- 1991 1995 1997
Government Ministry of Commerce and Ministry of Labor
Industry Korean Joint Committee for
(Strong) NGO Korean Federation of Small Businesses Migrant Workers
Positive Rationale Decrease of Labor Shortage Import of Foreign Workers as the Legal Status of the Worker
(Weak) Government NGO Ministry of Construction None Ministry of Finance and Economy Federation of Korean Trade Unions
Government Ministry of Justice, Ministry of Justice
(Weak) NGO Board of Economy and Planning, Ministry of Health and Society Federation of Korean Trade Unions None
Negative Rationale Disorder of Labor Market Increase of Wage
(Strong) Government Ministry of Labor Ministry of Trade and Industry, Small and Medium Business Administration
NGO None Korean Federation of Small Busin esses

Adoption of the program of
Results Adoption of the ITSF Working after Training for Foreigners
Source: Modified from Seol, 1999, p. 461.

Seols analysis is useful for understanding dramatic changes in the policy
positions of the major policy actors during the 1990s. Interestingly, the adoption of
the ITSF was a major watershed in dividing the position of the policy actors. For
example, the Ministry of Labor was against the importing of foreign workers at the
first stage of policy debate, but it changed its stance after the ITSF was adopted in
1991. On the contrary, small businesses, which argued for the necessity of the
importation of foreign workers in the late 1980s, opposed the enactment of the
import of foreign workers after the adoption of the ITSF. Despite its utility, it
should be noted that Seols analysis also faces similar criticism to that of Ko and
Lees Study (2004); it was not grounded on solid empirical data. Moreover, Seol
(1999) neglected to identify the complete picture of policy actors; as a result, some
important policy actors from the areas of political parties and the National
Assembly were not included at all.
To fully understand the evolution of Korean foreign labor policy, it is
necessary to appreciate the dissimilarities between the ITSF and the EPSF, which
were the chief immigration policy programs under consideration during the 1990s
and early 2000s. For purposes of comparison, their core characteristics are
summarized in Table 2.7. This table succinctly shows the core elements of each
policy option.

Table 2.7
Comparison of the ITSF and the EPSF
Adopted Time November 1991 August 2003
Relevant Act Immigration Control Act Act on the Foreign Workers Employment, etc
Ministry of Justice Ministry of Labor
Relevant Authority Committee on Foreign Industrial Labor Policy8 (Small and Medium Business Administration) Committee on Foreign Labor Policy (Established in Office for Government Policy Coordination)
Supporting Groups Business Associations Ministries for Industry and Small business NGOs for human rights Ministry of Labor
Labor Rights of Foreign Workers Not protected by labor law Protected by labor law
Visa Status of Foreign Workers Industrial Trainee (D-3) Non-professional Employment (E-9)
Maximum Length of Sojourn One year (in 1991) Two years (from 1993) Three years ( from 1996) Three years
aThe Committee on Foreign Industrial Labor Policy was abolished in August
2004. The Committee on Foreign Labor Policy established by the Act on the
Foreign Workers Employment, etc replaced its function.

This chapter consists of three major sections in which the chief theoretical
approaches relevant to this research are critically reviewed. The first section starts
with the review of policy network approaches, focusing on their main concepts and
typologies. It also deals with the limitations of the approaches and recent efforts to
address them. The review of the major streams of the policy network approaches is
expected to give some clues for developing network thinking and combining the
ideas with existing policy process frameworks.
The next section then addresses comprehensive policy process frameworks
based on network approaches. As network-based concepts and typologies show
only limited parts of network frameworks, reviewing comprehensive frameworks is
expected to provide implications for this dissertation. Among various policy
process approaches, two frameworks with policy network approachesi.e., the
Advocacy Coalition Framework (ACF), and the Organizational State framework
(OSF)are reviewed in detail to identify the framework of research in this area.

The last section focuses on the recent advancements of policy network
approaches. This section shows a variety of recent efforts to develop network-based
concepts and frameworks; thus, many studies of this section play critical roles in
making hypotheses of this thesis. Also, the literature on social capital and macro-
level factors, such as national culture, and political system, is reviewed in terms of
policy network. Specifically, social capital addresses the conditions for formulating
cohesive relationships among policy actors, and national culture shows the effects
of different cultural settings on the policy process. Considering that this research
analyzes a Korean policy case embedded in the political and cultural contexts
peculiar to Korea, the reviewing of cultural perspectives is expected to enhance the
research product by providing diverse counterpoints in understanding the policy
The Evolution of Policy Network Approaches
Network perspectives in the social sciences have a long history, developed
mainly from sociology and anthropology. German sociologist Simmel and French
anthropologist Levi-Strauss, the pioneers of the network approaches, suggested that
social processes are not the result of central steering or some kind of
prestabilized harmony but [emerge] through the purposeful interactions of
individual actors (Kenis & Schneider, 1991, p. 26).

Parallel with the development of social network analysis, research on
subgroups emerged. For example, Warner and Lunt already indicated that one of
the important goals of network analysis is to find subgroups or clusters in a
network by adopting the concept of clique (Warner & Lunt, 1941). The contribution
of the Manchester anthropologistssuch as Barnes, Mitchell, and Bottshould not
be neglected in the development of network analysis (Scott, 2000). Mitchell (1969)
maintained, for instance, that the quality of relations in networks can be measured
by such concepts as reciprocity, intensity and durability.
White and his colleagues developed a mathematical foundation for network
research. They focused on modeling of social structures, using social network
analysis. For example, White developed the method of block modeling to identify
subgroups in a network (Lorrain & White, 1971). Notably, social network analysis
gained popularity by Granovetters research in Getting a Job (1974), in which he
argued the strength of weak ties', i.e., new information is better acquired through
weak ties with those who are located in different work situations and have less
It was during 1990s that social network analysis attracted much attention in
the policy sciences (Carrington, Scott & Wasserman, 2005). Policy network
approaches, following the tradition of network perspectives in the social sciences,
tried to explain the policy process and policy outcomes in terms of the patterns of
relationships among policy actors. The common themes underlying policy network

approach were policy actors and their changing relationships (Borzel, 1998;
Carlsson, 2000; Jordan, 1990; Knoke, 1998; Marsh & Rhodes, 1992). The foci of
the approach, according to Carlsson (2000), lay in their (a) nonhierarchical way of
perceiving the policymaking process, (b) their focus on functional rather than on
organizational features, and finally (c) their horizontal scope (p. 505). In a similar
vein, many scholars focused on the network concepts as a replacement for a clear
divide of state and society (Atkinson & Coleman, 1989; Powell, 1990).
Major Concepts and Typology
In public policy studies, the idea of policy networks can be traced back to
Griffith (see Freeman, 1965) and Truman (1971), although their ideas were not
directly connected to other policy network studies (Jordan, 1990). According to
Griffith, studying the whirlpools of special social interests and problems was
preferable to studying formal institutions when it came to understanding real policy
processes (Freeman, 1965). Although Griffiths argument did not resonate instantly,
his idea has repeated itself over time in policy network approaches. That is, a
number of scholars focusing on the relationships among policy actors coined
similar terms, such as subsystem (Freeman, 1965), sub-government (Ripley &
Franklin, 1984), issue network (Heclo, 1978), and policy community (Richardson
& Jordan, 1979). Freeman (1965) proposed the concept of the subsystem, defined
as the pattern of interactions of participants, or actors, involved in making

decisions in a special area of public policy (p. 11). Similarly, Ripley and Franklin
(1984) defined sub-governments as the clusters of individuals that effectively
make most of the routine decisions in a given substantive area of policy,
comprising representatives of Congress, bureaucrats, and private groups that have
interests in the policy sector (p. 10).
These institutional network approaches cultimated in the popular concept
of the iron triangle, which focused on the solid relationships among
congressional committees, government agencies, and interest groups. In a similar
vein, the idea of the policy community focused on stable and close relationships
among policy actors. As the term community implies, the relationships in the policy
community are best characterized as closed village communities knitted together
by confidence, common calculations and specific climates (Kenis & Schneider,
1991, p. 29).
Rhodes (1985) detailed the concept of the policy community by outlining
some of its major characteristics, such as stability of relationships, continuity of
related membership, vertical interdependence based on shared service delivery
responsibilities, and insulation from other networks and invariably to the general
public (including national legislature) (p. 301). In contrast to the concept of policy
community, Heclo (1978) proposed the concept of issue network, in which
membership is wide open to those interested in the issues under consideration. The

issue network is distinguished from other network concepts in that it implies the
policy process can be dominated by an unpredictable and complex network.
The emergence of diverse concepts based on network ideas prompted the
introduction of typologies of policy networks (Atkinson & Coleman, 1989; Benson,
1982; Rhodes, 1990). Typically, Rhodes (1990) furthered the idea of diverse
structures in policy networks by suggesting a typology comprising five types of
networks based on membership, interdependence, and resources: i.e., policy
communities, professional networks, intergovernmental networks, producer
networks, and issue networks (See Table 3.1). In his typology, for example, a policy
community was characterized by stable and very limited membership, vertical
interdependence, and limited horizontal articulation.

Table 3.1
Types of Policy Networks: The Rhodes Model
Type of network Characteristics of network
Policy community/ Territorial community Stability, highly restricted membership, vertical interdependence, limited horizontal articulation
Professional network Stability, highly restricted membership, vertical interdependence, limited horizontal articulation, serves interest of profession
Intergovernmental network Limited membership, limited vertical interdependence, extensive horizontal articulation
Producer network Fluctuating membership, limited vertical interdependence, serves interest of producer
Issue network Unstable, large number of members, limited vertical interdependence
Source: Rhodes and Marsh, 1992, p. 14.
Despite the development of the concepts and typologies on the policy
networks, scholars in the field have not shown convergence on the definition of
policy network (Waarden, 1992), which is a major barrier for the network
approaches to expand their territory. Nevertheless, some common features of the
concept suggested by scholars may play a considerable role in recognizing the
contours of the idea (Borzel, 1998; Carlsson, 2000). According to Carlsson (2000),
for example, the common components underlying networks include actors,
linkages, and boundary. Carlsson (2000) explains:

A policy network is described by its actors, their linkages and its boundary.
It includes a relatively stable set of mainly public and private corporate
actors. The linkages between the actors serve as channels for
communication and for the exchange of information, expertise, trust and
other policy resources. The boundary of a given policy network is not in
the first place determined by formal institutions but results from a process
of mutual recognition dependent on functional relevance and structural
embeddedness (p. 505).
Borzel (1998) also detailed the common features in the policy network
approaches. He viewed a policy network as a set of relationships characterized as
stable and non-hierarchical in nature, in which policy actors with common policy
interests try to achieve their goals by exchanging their resources.
Formal Network Analysis
The utility of formal network analysis springs from the uniqueness of the
relational data. Relational data are not about the traits of any specific actor; rather,
they measure the ties or connections among actors. Therefore, it is difficult to
understand relational data by adopting conventional variable analyses. To overcome
this problem, the social network analysis (SNA) focuses on relations among actors
and provides a variety of analytical tools that are useful for analyzing the
characteristics of the relations (or structures).
Graph theory supports the theoretical backgrounds of the SNA. In the
graph theory, the two most fundamental elements are points and lines. A relational

matrix can be described as a graph of points connected by lines (Scott, 2000, p.
64). Any points that are directly connected with one another are termed
neighborhood, and the total number of adjacent points indicates degree of
connection. The distance between points can be measured by calculating the length
of the shortest path between them (Scott, 2000). The following basic concepts of
the SNA provide more information about the SNA ideas and their characteristics.
In describing a network, one of the most basic elements is the connection
of the network. To measure characteristics of connection, scholars in the area of the
SNA have developed several concepts, such as density, connectivity, and
reachability. The concept of density in a binary network can be defined as the
proportion of all possible ties that are actually present (Hanneman & Riddle, 2005,
in Chapter 7). Given that more connected individuals may be able to access more
resources and information, density can play a significant role in measuring social
capital in a given network. Point connectivity is measured by calculating the nodes
to be removed in order for one actor not to reach another actor. The concept of
connectivity is significant in that it implies the degree of dependency of an actor to
the other in the network. Reachability measures whether an actor can be connected
to another, not considering how many actors go between them. The concept of
reachability provides a clue about whether or not a network is divided and, if so,

Another fundamental element in a network is the distance among actors.
The concepts around connection are mainly about the existence and the degree of
relations among actors, whereas the concept of distance implies the strength of ties.
One of the most popular measurements of distance between actors is the geodesic
distance defined as the number of relations in the shortest possible walk from one
actor to another (Hanneman & Riddle, 2005, in Chapter 7). Derived from this
concept, the geodesic path refers to the shortest path between actors.
The ratio of ties within groups to ties between groups provides good
indications about the network structure. The E-I (external internal) index,
developed by Krackhardt and Stem (1988), is the number of ties to outsiders from
group members, subtracted by the number of ties to insiders from the same group
members and divided by the total number of ties. The range of the E-I index is from
-1 (totally internal orientation) to 1 (totally external orientation). The index can be
measured at three different levels: the levels of the whole network; groups; and
individuals. This index is useful for indicating the degree of openness of groups and
the embeddedness of individuals in their group.
Some basic characteristics of the SNA should be described in detail to
provide a complete understanding of the SNA. First of all, relational data are
usually described by square-shaped matrix arrangements (adjacency matrix), in
which both the rows and columns share the same actors, or cases, or events;
attributes data, in contrast, adopt matrices with actors by properties (variables).

Adjacency matrices can be described either by undirected data or by directed data.
Undirected data are simple and relatively easy to handle. But directed data are
useful for studying power or influence among actors as they provide information
about the direction of the exercise of power or influence (Scott, 2000). Relational
data can be measured either by binary or by valued types. Binary-type data have a
binary digit, 1 or 0, that shows the presence or absence of a specific relation
between actors. Valued data are different from binary-type data in that the degree of
relationship is expressed by an actual value. The strength of relations can be better
understood by valued data. The levels of analysis in SNA primarily depend on the
research questions to be answered.
Second, there are two major approaches to relational dataego-centric and
socio-centric networks. Egocentric networks focus on the individual, rather than on
the whole network, thereby providing better information about local networks of
the focal actor. In contrast, socio-centric approaches are superior for recognizing
the overall structure of the network. Combining these two approaches, one can
obtain an improved picture of how individual actors selections are connected with
the structure of the whole society (or system). In a similar vein, the contrast
between total networks and partial networks by Mitchell (1969) shows the
limitations of the network approaches. According to Mitchell (1969), the total
network of a society lies in the general ever-ramifying, ever-reticulating set of
linkages that stretches within and beyond the confines of any community or

organization (p. 12). However, the SNA in a real world situation inevitably has to
narrow its focus by selecting specific aspects of the network. Although multiplex
relations in a network can be described in a matrix or in a graph, it is almost
impossible to contain all the possible aspects of a network in a matrix.
Third, network analysis demands a census rather than a sampling in
designing research. As each actor in a population has his or her own relationships
with other actors, the actors in the population are not independent. Moreover, they
are not interchangeable, which renders sampling useless. Although some network
studiessuch as egocentric network researchdo not demand information about
the entire network, the holistic approaches face difficulty in adopting conventional
sampling methods.
With the basic concepts of the SNA, many scholars developed more
complicated and significant methods for studying social activities. First, Krackhardt
(1994) introduced the concept of hierarchy, which captures a network structure
based on vertical differentiation. To measure the degree of hierarchy, he adopted an
ideal hierarchy in the form of an out-tree graph, in which all actors are connected
and have only one higher actor except the leader. The ideal type of hierarchy
consists of four separate measures: connectedness, hierarchy, efficiency, and least
upper bound. These measures are based on the following assumptions: A network is
more hierarchical (1) when all of the actors are connected in the same component;
(2) when the network has fewer reciprocated ties; (3) when each actor has only one

person in authority; and (4) when each pair of actors shares an actor that has direct
ties to both members of the pair (Krackhardt, 1994). Despite the significance of
network structure in terms of vertical differentiation in understanding the policy
process, the degree of hierarchy of policy networks is not easy to measure. In this
respect, Krackhardts concept of hierarchy provides a good tool for measuring
Second, another aspect introduced by social network analysts lies in the
relations between network positions and social roles. In this perspective, the focus
is the patterns of interaction among actors, not the attributes of the actors. Those
who share similar relational patterns face similar opportunities and constraints,
thereby playing similar social roles in the network. In measuring similarity or
equivalence of patterns of relations, three major approaches have been put forward:
structural equivalence, automorphic equivalence, and regular equivalence
(Hanneman & Riddle, 2005). The concept of structural equivalence addresses the
extent to which actors share the same relationships to other actors; i.e., two actors
are structurally equivalent if they have exactly the same relationships to all other
actors. Structural equivalence can be measured in many ways, including CONCOR
and TABU search protocols (Scott, 2000). Both of them try to find the smallest sum
of within-block variances of the ties among actors. Automorphic equivalence
spotlights a set of actors that can be identified as a sub-group, whereas structural
equivalence focuses on dyadic relations. Automorphic equivalence seeks to

measure the extent that a set of actors in a sub-group is substituted by a set of actors
in other sub-groups. Finally, the concept of regular equivalence emphasizes the
relations to other regularly equivalent actors. Two actors are considered as being
regularly equivalent when they are connected to equivalent others in the same way.
A precondition for these equivalence concepts is the development of measuring
equivalence or similarity. For this, a variety of tools, such as Pearson correlation
coefficients and Euclidean distance, were devised. Pearson correlation coefficients
are useful for valued data. The range of Pearson correlation coefficients is from -1
to 1, depending on the degree of structural equivalence between actors. Euclidean
distance is not sensitive to the linear association and is useful for both valued and
binary data. The degree of similarity of the actors can be visualized by using a
dendogram. Also, multi-dimensional scaling tools are useful for mapping the actors
on a multi-dimensional space, revealing diverse aspects underlying the similarities
of actors.
Lastly, network approaches provide insight into how to approach and
measure the concept of power, one of the most cited concepts in the social sciences.
Despite its prevalence in studies, however, the measurement of power is often
controversial. Social network analysis adopts the concept of centrality to measure
power, using such indices as degree, closeness, and betweenness (Hanneman &
Riddle, 2005). The concept of degree is meaningful for measuring power with the
assumption that more ties usually tend to indicate more power. Along with this idea,

Bonacich (1987) suggested that centrality should be measured in terms of the
number of connections between the neighborhood actors as well as ones own
number of ties. Ones power increases when one has more ties with other actors and
when those other actors have fewer ties. The concept of betweenness shows another
facet of power. A popular way of measuring betweenness is to calculate the extent
to which an actor falls on the shortest way between policy actors (Hanneman &
Riddle, 2005). Another way of measuring the concept of centrality is using the idea
of closeness. Given that distance between actors is important in determining the
degree of power, it is reasonable to measure the closeness between actors for
Limitations of Policy Network Approaches
Despite the accelerated development of policy network analysis, the policy
network perspectives have limitations as well (Adam & Kriesi, 2007, Dowding,
1995). First, critics point out that policy network approaches lack explanatory
power (Blom-Hansen, 1997; Dowding, 1995; Klijn & Koppenjan, 2000; Marin,
1990; Peters, 1998). If policy networks are to be effective predictors of policy
outcomes, they argue, the features of networksnot just the characteristics of
individual policy actors in the networksshould have sufficient explanatory power.
For this reason, policy network perspectives have been criticized as being more
useful as metaphors than as systematic models (Dowding, 1995). In this respect,

Peters (1998) also indicated there is a tendency in the literature for networks to be
the dependent variable for other systemic changes, rather than an important
explanatory factor (p. 25). Also, given that the relationships among policy actors
are not fixed over time, a network analysis based on a specific period cannot reveal
the overall picture of the network. In particular, in cases where the boundaries of
networks are fluid over time, the structures of networks can be significantly
changed, depending upon the point in time when the data are collected.
Second, others argue that the policy network approaches lack critical
theoretical bases and cannot stand alone without incorporating macro and micro-
level analyses (Blom-Hansen, 1997; Dowding, 1995; Klijn & Koppenjan, 2000;
Rhodes & Marsh, 1992). According to Blom-Hansen (1997), the absence of a
model of the actor in the policy network perspectives makes it difficult to answer
such questions as why networks exist and why they change. As the concept of
policy networks that focuses mainly on meso-level analysis, it should be
supplemented by macro-level theories, such as theories of the state (Rhodes &
Marsh, 1992).
Third, there is no shared definition of policy networks among the scholars
in this field (Borzel, 1998; Jordan, 1990; Jordan & Schubert, 1992; Waarden, 1992).
For example, Waarden (1992) identifies eleven different definitions of policy
networks and seven dimensions in the existing policy network literature. As
definitional agreement on major terms is the starting point of focused research, this

lack of consensus is problematic to cumulate research outcomes. Moreover, in the
case of analyzing networks among organizations, which are major actors in the
policy process approaches, many additional problems emerge in adopting the SNA.
For example, it is not apparent who should be sampled to represent the
relationships among organizations in the network analysis. Also, it is difficult to
measure the frequency of communication between organizations because
organizations consist of a number of real actors.
Fourth, Peters (1998) points out that the linkage between policy network
approaches and policy-process models is not clear. That is, the effects of differently
structured networks at each stage of the policy process should be elaborated with
empirical evidence. Peters (1998) argued, (t)he nature of networks should have a
great deal to say about the opening and closing of policy windows as well as
about how issues are constructed in order to make them more suitable for
institutional agendas. Unfortunately, that linkage is rarely made explicit by network
theorists (p. 26).
Lastly, despite the large efforts at collecting and analyzing relational data,
the results from the network analysis do not provide significant findings (Dowding,
1995). This criticism is closely connected with controversy on the appropriate
targets of network analysis. Critics argue that the core aspects of the policy
processe.g., political power, norms, and rules embedded in the policy
networksare difficult to capture by network analysis.

Policy Process Frameworks with Policy Network Perspectives
Network perspectives have been incorporated into some recent policy
process theories and frameworks. In many cases, these theories and frameworks
only borrowed basic ideas from network perspectives. For example, the Punctuated
Equilibrium Theory suggested the concept of policy monopoly, implying that some
limited number of policy actors dominate the policy process in a subsystem
(Baumgartner & Jones, 1993). This concept can be regarded as a revised version of
the concept of iron triangles, in the sense that both of them underline stable
relationships among the actors in a policy subsystem.
Also, Kingdons Multiple Streams Approach employed the concept of
policy community in explaining its policy streams. According to Kingdon (1995),
policy alternatives experience the process of softening up to obtain their legitimacy
in a given policy community. Zahariadis and Allen (1995) developed the concept of
policy community in the multiple streams approach by examining how the
structures of policy communities influence policy innovation. Focusing on the
integration of ties among policy participants in terms of size, mode, capacity, and
access, they then argued that differently structured networks have different
trajectories of alternatives: Less-integrated networks are more likely to facilitate a
quantum to gradualist evolution and more-integrated networks are more likely to
follow an emergent to convergent pattern (Zahariadis & Allen, 1995, p. 75).

It should be noted, however, that some policy process frameworks more
fully adopted the network perspectives. This section focuses on these frameworks
by reviewing the Advocacy Coalition Framework (ACF), and the Organizaitional
State Framework (OSF).
Advocacy Coalition Framework
The ACF is a conceptual framework that attempts to explain policy changes
and learning within a policy subsystem, typically over a decade or more timeframe.
Sabatier and Jenkins-Smith first introduced this framework in 1988. Since then,
more than 100 research papers have dealt with the framework (Sabatier & Weible,
2007). Many of them focused on U.S. energy and environmental policy, mainly in
the context of the U. S. However, recently the ACF has been applied to a variety of
policy domains in diverse national configurations to investigate its applicability
(e.g., Sato, 1999).
According to Sabatier and Weible (2007, pp. 191 192), the logical
departure points of the ACF can be summarized as the following three foundation
(1) a macro-level assumption that most policymaking occurs among
specialists within a policy subsystem but that their behavior is affected by
factors in the broader political socioeconomic system; (2) a micro-level
model of the individual that is drawn heavily from social psychology; and
(3) a meso-level conviction that the best way to deal with the multiplicity of
actors in a subsystem is to aggregate them into advocacy coalitions.

Basically, the ACF stresses shared belief systems as the cohesion of coalitions, the
role of information and learning as a motivator of policy change, and the role of the
policy broker as a mediator among conflicting coalitions (Sabatier & Jenkins-Smith,
1999; Schlager & Blomquist, 1996). Also, in explaining policy changes, the ACF
focuses on policy elites, especially on their belief systems. In other words, policy
elites, who are assumed to have a hierarchically ordered set of beliefs, the ability
to process information, and a set of goals, or preferences, participate in policy
processes to translate their belief systems into policy outcomes (Schlager, 1999, p.
A second feature of the ACF lies in its unit of analysis, a policy subsystem.
According to Sabatier and Jenkins-Smith (1993), the best unit of analysis in
studying policy changes is not any formal institution but rather a policy subsystem.
The concept of policy subsystem is similar to that of issue networks, in the sense
that both of them focus on a specific substantive policy issue and emphasize the
openness of their network membership to a variety of policy actors (e.g., all levels
of government, researchers, and the media). As the ACF employs the policy
subsystem as a unit of analysis, it can provide a variety of group dynamics in the
policy process (Schlager & Blomquist, 1996). Along those lines, the ACF proposes
the concept of advocacy coalitions as a tool for analyzing the behavior of hundreds
of policy actors within a policy subsystem.

Focusing mainly on the policy changes over a decade or more, the ACF
emphasizes the interaction among competing coalitions in a policy subsystem.
More specifically, the ACF suggests four major paths to policy changes: (1) policy-
oriented learning based on scientific and technical information; (2) the changes of
external (system) events, such as socioeconomic conditions, public opinions,
systemic governing coalitions, and policy decisions and impacts from other
subsystems; (3) internal shocks such as disasters that have occurred from within a
policy subsystem; and (4) negotiated agreement emerging under such
circumstances of a hurting stalemate (Sabatier & Jenkins-Smith, 1999; Sabatier &
Weible, 2007).
Also, the ACF underlines the roles of relatively stable parameters and
external events as constraints and resources of subsystem actors. Recently, the
framework made a new additioncoalition opportunity structureto its
framework to broaden its applicability to other political contexts (e.g., corporatist
regimes). Figure 3.1 displays the conceptual scheme of the ACF.

Source: Sabatier and Weible, 2007, p. 202.
Figure 3.1. Diagram of the Advocacy Coalition Framework (2005 version).

To attain a complete understanding of the ACF, some of its core concepts
such as advocacy coalitions, belief systems, and policy subsystemsshould be
detailed. First, the ACF shows that the unit of advocacy coalitions is very useful for
analyzing the behavior of hundreds of policy actors within a policy subsystem.
According to Sabatier and Jenkins-Smith (1999), an advocacy coalition consists of
policy actors from private sectors as well as governmental organizations that both
(1) share a set of normative and causal beliefs and (2) engage in a nontrivial degree
of coordinated activity over time (p. 120). From the policy-network standpoint,
the advocacy coalition is in line with the issue network rather the policy community
because it underscores openness of networks in which a wide range of policy actors
can participate in the policy process.
Second, the ACF assumes that policy actors intend to translate their beliefs
into policy outcomes. The ACF assumes that policy elites have relatively stable
belief systems. According to Zafonte and Sabatier (2004), policy elites have well-
integrated policy belief systems that link fundamental substantive and distributional
values, perceptions of the severity and causes of policy problems, and perceptions
of the proper approaches to be used in addressing these problems (p. 77). Thus,
the ACF posits a reasonable explanation for the stability of the coalitions. In the
ACF, belief systems are made up with three tiers of beliefsdeep core beliefs,
policy core beliefs, and secondary aspects of beliefs. According to Sabatier and
Jenkins-Smith (1999), deep core beliefs are defined as basic ontological and

normative beliefs, such as the relative valuation of individual freedom versus social
equality, which operate across virtually all policy domains, policy core beliefs are
a coalitions basic normative commitments and causal perceptions across an entire
policy domain or subsystem, and finally secondary aspects of beliefs relate to the
seriousness of the problem or the relative importance of various causal factors in
specific locales, policy preferences regarding desirable regulations or budgetary
allocations, the design of specific institutions, and the evaluations of various actors
performance (pp. 121-122).
A similar concept related to the belief systems is policy core policy
preferences. The concept is defined as beliefs that (i) are subsystemwide in scope,
(ii) are highly salient, and (iii) have been a major cleavage for some time (Sabatier
& Jenkins-Smith, 1999, p. 134). Parallel with this concept, the ACF emphasizes
policy core beliefs more than the other levels of beliefs as a core determinant
influencing coalition members behavior.
Third, another significant issue in the policy process theories is the
dynamics in policy changes. According to Sabatier and Jenkins-Smith (1999), the
ACF accounts for three factors in explaining policy changes: (1) the interaction of
competing coalitions within policy subsystems, (2) the effects of relatively stable
parameters on the constraints and resources of subsystem actors, and (3) the
changes of external (system) events, such as socioeconomic conditions, public
opinions, systemic governing coalitions, and policy decisions and impacts from

other subsystems (p. 149). Also, the ACF indicates that belief changes can lead to
policy changes and that belief changes occur through either policy-oriented
learning or turnover of policy elites.
Some critics have pointed out the limitations of the ACF (deLeon, 1994;
Hann, 1995; Schlager & Blomquist 1996; Zahariadis, 1995). First, Maloney et al.
(1994) argued that the ACF failed to distinguish more important policy actors from
less important ones in a given policy area because it neglected the distinction
between insiders and outsiders in coalitions. In addition, they also maintained that
the ACF did not account for the phenomenon that a policy domain might be
structured by harmonious and stable relationships among participants, as the
concepts of iron triangle and policy community demonstrated (Marsh, 1998;
Rhodes, 1990).
Second, the ACF neglects collective action problems (Olson, 1965;
Schlager, 1995). Despite the distributional issues within a coalition, the ACF
assumes the existence of highly coordinated behavior among the coalition members.
As Olson (1965) indicated, however, common interests do not necessarily lead to
collective action; rather, the members of a coalition might experience internal
conflicts in deciding their collective action in case they do not share the same
benefits or the same costs. Also, in cases where the actors in a coalition cooperate
over time, the ACF neglects the conditions that lead actors in a coalition to
coordinate their activities. As some scholars have indicated, however, coordinated

activities among policy actors may require such preconditions as shared norms,
mutual trust, respect, and interdependency (Fenger & Klok, 2001; Kenis, 1991;
Paxton, 1999; Putnam, 1995).
Third, although the ACF may be useful in describing policy change after the
fact, it lacks the ability to pinpoint when policy change actually takes place or how
advocacy coalitions develop (Zahariadis, 1995). Sabatier and Jenkins-Smith (1999)
maintained that information and learning were effective tools to change the
secondary aspects of beliefs and that major policy changes stem from events
outside the policy subsystem. However, the ACF cannot explain what kinds of
environmental changes are critical for a specific policy change or how those
changes influence policy outcomes. As a result, the ACF is criticized for discarding
the core mechanisms of policy change into a black box (Caimey, 1997; Sato,
Fourth, the concept of belief systems has some drawbacks. These
limitations need to be dealt with in detail because one of main purposes of this
dissertation is to amend these limitations. Above all, it should be noted that it is
very difficult to identify other policy actors belief systems accurately and that,
even if it is possible, the process may require too much time and too high costs to
manage. Also, given that an organization consists of various individuals who often
do not share the same belief systems, their argument should be complemented by
additional evidence that policy actors try to interact more with those who have

similar belief systems in the target organization. Moreover, the belief systems
argument disregards the critical roles of organizational actors in the policy process.
In other words, in the cases of organizational actors, a serious question emerges:
Who should represent the belief systems of the organization? The answer may be
very different depending on the institutional settings of the organization under
consideration.19 For example, in a hi era rchical organization, the policy core beliefs
of the organizational actor may be likely to resemble those of the highest decision-
makers in the organization, whereas in a more horigental organization, one should
expect higher variance among the organizations members.
Another problem of the concept of belief systems comes from the term
belief. The term implies that policy actors should be human beings, not robotic
organizations because it seems awkward for organizations to contain any coherent
belief systems. Also, Weible and Sabatiers explanation about the deep core beliefs
supports this argument: deep core beliefs are exogenous to a policy subsystem,
developed during childhood, and very resistant to change (Weible & Sabatier,
2005, p. 196). However, considering the significant roles of organizations in the
policy process, organizations should not be ignored in analyzing the policy process.
A question then emerges: how can the belief systems of organizations be
formulated and measured?
19Scharpf (1997) identifies several institutional settings and tries to combine them with
specific modes of interaction.

Finally, the concept of belief systems entails the problem of measurement.
Although Sabatier and Jenkins-Smith (1999) argued that questionnaires and content
analyses of documents and testimonies are useful for identifying belief systems, the
measurement of belief systems can be problematicespecially considering the
indication that some of the ACFs major terms have unclear definitions (deLeon,
1994). As deLeon (1994) noted, such terms as core beliefs, secondary beliefs and
the level of conflict are undetermined prior to observing the phenomenon, a
condition that undercuts ACFs predictive elements. Also, notably, it is difficult to
distinguish policy elites beliefs from the policy interests of the organization they
belong to because most policy elites are apt to follow the policy preferences
supported by their affiliations.
Organizational State Framework
Laumann and Knoke (1987) introduced the Organizational State
Framework (OSF) to increase the apprecialtion of the policy process, based on a
policy network perspective. This framework started when the researchers applied
the framework to the U.S. energy and health policy domains (Laumann & Knoke,
1987), and then it extended its applicable areas to labor policy domains across the
U. S., Germany, and Japan (Knoke et al., 1996). The OSF views the modern
industrial polity as a complex of formal organizations in conflict with one another
over the collective allocation of scarce societal resources and considers the state as

an autonomous social formation whose strategies emerge from the basic
organizational imperatives of coping with environmental uncertainties, resource
scarcities, and socio-legal constraints (Laumann & Knoke, 1987, p. 8).
Basically, the OSF focuses on such concepts as policy interests, coalitions,
and social networks, combining them into a distinctive framework for analyzing
national policy domains. The following argument shows the gist of this framework:
Policy players constantly importune instrumental ties that they believe will
further their policy interests. However, these alliances are channeled by
embedded patterns of social networks. These linkages are not randomly
patterned; rather, they are hemmed in by past histories of collaboration
and antagonism, by current conditions of trust and distrust, and by
anticipation of future costs and benefits from available alternatives (Knoke
etal., 1996, p.219).
The unit of analysis of the OSF is the policy domain. According to Knoke
and Laumann (1982), policy domain is a substantively defined criterion of mutual
relevance or common orientation among a set of consequential actors concerned
with formulating, advocating, and selecting courses of action (i.e., policy options)
that are intended to resolve the delimited substantive problems in question (p. 256)
Laumann and Knoke (1987) argued that contemporary policy domains are
composed of two processes: the process of determining participants and that of
framing issues to which the participants try to influence. Policy domains comprise
policy actors who share common policy interests, although their policy preferences

may not be identical, and recognize the necessity of considering other actors to
realize their policy goals.
Knoke et al. (1996) suggested some major components of the policy domain
framework, including policy actors, policy interests, power relations, collective
actions, and jointly occupied positions. They then argued that the configurations
of these components at any time provide policy analysts with comprehensive maps,
or multidimensional snapshots, of the domains social structures and activities
(Knoke et al., 1996, p. 10). Table 3.2 illustrates some typical examples of the

Table 3.2
Components of the Policy Domain Framework
ORGANIZATIONAL ACTORS Interest Groups Peak Associations Government Institutions Unions; trade groups; PIGS; corporations; bureaucracies Federations; chambers; quasi-official advisory bodies Legislatures; courts; ministries; regulators; official advisory councils
POLICY INTERESTS Subfields Issues Events POWER RELATIONS Broad areas of focal concern; collective bargaining; markets General substantive matters; wages, hours, conditions Legislative bills; court suits; regulatory proclamations
Information exchange Resource exchange COLLECTIVE ACTIONS Ideas, data, strategies, advice Funds, facilities, votes, coercion
Mobilization Publicity Lobbying Coalition building; social movements Mass media blitzes; targeted mailings Contacts with government officials; legal suits
JOINTLY OCCUPIED POSITIONS Issue publics Event publics Advocacy circles Action sets Shared profiles of interests in all domain issues Interests in a specific event, regardless of outcome preferences Shared preferences for a specific event outcome Active collaboration to produce favorable event outcome
Source: Knoke et al., 1996, p. 11

The OSF operates on two assumptions: (1) corporate actors are key actors
in state decision-making; and (2) supraindividual structural arrangements should be
considered for explaining policy domain event participation (Laumann & Knoke,
1987, p. 9). In this perspective, major policy actors are formal organizations that
attempt to realize their policy goals in the policy process. That is, the major players
in the OSF are formal organizations who have (1) variable interests in a range of
issues in a national policy domain and (2) relevant mobilizable resources
(Laumann & Knoke, 1987, p. 5, emphases in original). Policy actors pursue their
policy interests through interorganizational alliances in which policy actors seek
trading off their support on events of lesser interest in return for help on events of
higher concern (Knoke et al., 1996, pp. 8-9). The OSF emphasizes the roles of
information (both technical expertise and political knowledge) and resources (both
material and symbolic) in accessing key positions in the policy process (Knoke et
al., 1996, p. 8). Accordingly, policy outcomes are considered as dependent upon the
actors interests and resources (Laumann & Knoke, 1987).
Knoke et al. (1996) detailed the OSF through a causal model in which
organizational reputations and policy activities were considered to be influenced by
organizational characteristics such as policy interests, resource capacities, and
organizational types via communication and support networks (see Figure 3.2).

Policy Interests

^ Support
Source: Knoke et al., 1996, p. 106.
Figure 3.2. Causal Model of Relations among Organizational Characteristics,
Exchange Networks, and Reputations and Event Activities.
Overall, the model emphasizes the intensity of policy interests and
resources as critical factors in deciding policy outcomes. The perspective then
considers information and resource exchange networks as basic political
structures that affect power relations among policy actors. The model emphasizes
the role of policy interests in the entire policy process, assuming the variable of
policy interests to be formulated through an organizations long-standing
commitments to specific substantive concerns (Knoke et al., 1996, p. 105).
Resource capacities include symbolic as well as material advantages by which
organizations can enhance their accessibility to policy-making and their influence
on policy outcomes. The variable of organizational type distinguishes among

governmental actors with formal policy-making authority, business associations,
and public interest groups (Knoke et al., 1996, p. 105). These organizational
attributes affect decisions regarding the organizations locations in both
communication and support networks. Although the organizational attributes do
not totally determine the organizations centralities in the networks, a principle is
suggested to obstain a better understanding of the linkages between
organizational characteristics and locations in the networks: Actors having
broader policy interests, larger resource capabilities, and public authority can
more easily convert those advantages into proximity to the centers of each
exchange network (Knoke et al., 1996, p. 105).
Finally, the OSF causal model depicts the relations of the exchange
networks to organizational reputations and policy activities with the following
hypothesized principle: The more central an organizations location in an
exchange network, the higher its reputation and the more extensive its activities
(Knoke et al., 1996, p. 107). The results from empirical tests of this model prove
the effectiveness of this model. Knoke et al. (1996), for example, tried to apply the
OSF to the comparative labor policy domains of the U. S., Germany, and Japan.
They found, in this particular policy domain, that the more central an
organizations locations in both the communication and the support networks, the
higher its reputation for public policy influence and the greater its public policy
event activities (Knoke et al., 1996, p. 121).

Regarding the issue of collective action in the policy process, Knoke et al.
(1996) maintained that it is facilitated when policy actors share common goals and
create ties among them. To explain the dynamics of collective action, the OSF
adopts the concepts of advocacy circles and action sets. According to Knoke et al.
(1996), an advocacy circle encompasses three or more formal organizations within
an event public who communicate directly or indirectly among themselves about
policy matters and who prefer the same outcome on that event (Knoke et al., 1996,
p. 21). This concept emphasizes the same policy preference among the members
but does not take into account the coordinated actions among them. The matter of
coordination is characterized by action sets, in which the members not only share
policy preferences but are also linked together to affect the policy outcome by
collective actions.
Along with these concepts, Knoke et al. (1996) tried to identify the factors
predicting an organizations participation in the action sets. They hypothesized that
the main predictors are (1) the breadth and intensity of an actors interests over the
range of policy domain issues and (2) the availability of organization resources
financial and material assets, influence reputations, network locations (Knoke et
al., 1996, p. 138). Their empirical tests for these hypotheses indicate that positive
relationships existed between policy interests and action-sets participation in all
three of the nations being analyzed (Knoke et al., 1996).

Collective action is embodied by policy actors strategies. Some of the
most popular tactics adopted by organizations in labor policy domains of the U.S.,
Germany, and Japan were identified as the following:
Participation in formulating this proposal or suggested alternatives;
Formal contacts with government officials, legislature, or courts, such
as testifying at hearings, serving on commissions, or filing friend of
the court briefs;
Informal contacts with important policy actors such as government
officials, legislature, or non-governmental organizations;
Using mass media to publicize an organizations opinion about the
event; and
Mobilizing members of the organization or the general public to
influence the public authorities (Knoke et al., 1996, p. 148).
No single framework is without its limitations. Although the OSF provides
many useful concepts, critics have pointed out its weaknesses as well. First, Alford
and Friedland (1985) focus on the managerial perspective of the OSF, in which
society is described as dominated by organizations, and the state is the central
organization (Alford & Friedland, 1985, p. 171). The OSF assumes that
organizations, not individuals or classes, are key players in the whole policy
process. Critics argue that the behaviors of an organization are often greatly
influenced by specific individual or class interests (Alford & Friedland, 1985; Scott,

1995). In this sense, without serious consideration of this aspect, real mechanisms
in the policy-making process may not be clarified. For example, individuals such as
politicians, academics, and journalists are often crucial policy actors, but their
behavior in the policy process can be better understood when their individual
beliefs, attitudes, and policy preferences are analyzed rather than those of their
organizations. In other words, it is not rare to detect politicians who do not stand by
the same policy position supported by their party.
The second criticism is that the OSF lacks the ability of considering the
significance of institutional arrangement. It is well recognized that political
contexts make significant differences in the policy-making process. However, the
OSF does not articulate the roles of history, culture, or institutions in its approach.
Nevertheless, the OSF tries to ascribe the differences in the patterns of policy
networks among nations to each nations peculiar historical, cultural, or
institutional backgrounds (Knoke et al., 1996). Given the importance of contextual
factors at the level of an organization or a state, these factors should be located
explicitly in the OSF. In this sense, Boli (1997) also argued for the significance of
cultural analyses in the policy process: To make sense of this plethora of
information, though, we need more than the details. We also need cultural analyses
in which the assumptions and worldviews of mobilizing interests and legitimated
authorities are not treated as givens but as data (p. 187).

Last, a relatively minor criticism is the paucity of OSF-based research.
Admitting that a framework develops through a variety of empirical tests, it is a
virtue for a framework to attract much research as a way developing or sharpening
its conceptual framework. In this respect, however, the OSF fails the criterion
because so little research has adopted the perspective to study the policy process.
As Gould (1997) argued, a major explanation of this condition might lie in the
problems of operationalization: This slack between theoretical propositions and
their operational counterparts makes it difficult to know when to be surprised by
the findings and when not to (p. 1746)
Recent Advancement of Policy Network Approaches
To overcome the drawbacks of the early policy network approaches,
scholars have recently devoted considerable efforts in a number of areas: some
scholars tried to explain collective action in the policy networks; some others
examined the effects of network structures on policy outcomes', some others focused
on social capital in the policy network approaches', and still others showed interest
in combining macro-level factors with policy networks approaches.
Collective Action among Policy Actors
Collective action among coalition members must be fully understood if the
concept of a coalition is to be effective for understanding the policy process. In this

sense, it is meaningful to review some major studies on collective action or
interaction in policy networks. It is a good starting point to review possible
relationships among policy actors in a coalition. As to the relationships among
coalition members, Bish (1978) already suggested four kinds of relationships
among independent political units based on public-choice theory: cooperation,
collusion, competition, and coercion. According to Bish (1978), cooperation is the
joint action by two or more parties for mutual benefit (p. 22). Collusion is similar
to cooperation in that it provides benefits to the actors involved, but the effect on
third parties is very different; i.e., collusion entails third-party costs, whereas
cooperation does not. The concept of competition indicates rivalry between two or
more parties for the favors of a third party, or in some cases for the capture of an
object for which no party has a proper right (Bish, 1978, p. 23). Finally, coercion
implies a relationship that the coerced party has no other choice but to follow the
preferences of the coercing party to avoid any sanction from the party. These
concepts can be a basic foundation for understanding the relationship among policy
Regarding the formation of policy networks, Konig and Brauninger (1998)
maintained that policy actors choice of network contacts is greatly influenced by
the similarity of their preferences.20 The y also argued that institutional settings
20 Kon ig and Brauninger (1998) viewed preferences as consisting of two factors: actors
interest and their policy position on political events (p. 447).

play a significant role in deciding policy actors network contacts. Policy actors
tend to mix with policy decision-makers who have formal voting power; those who
are in the same sector tend to exchange resources in the policy process (Konig &
Brauninger, 1998). Also, the sectoral view on institutional settings focuses on the
policy actors embeddedness in a social system. In this view, policy actors in the
same sector make more contacts with one another than with those in different
sectors (Coleman, 1990; Konig & Brauninger, 1998; Laumann & Knoke, 1987).
Many scholars have studied factors or conditions for collective action
(Fenger & Klok, 2001; Klijn & Koppenjan, 2000; Knoke et al., 1996; Konig &
Brauninger, 1998; Scharpf, 1978; Zafonte & Sabatier, 1998). Resource-dependence
approaches emphasize that interactions among policy actors are facilitated by the
necessity of accessing the resources of other policy actors (Aldrich, 1976; Yuchman
& Seashore, 1967). Based on these approaches, Scharpf (1978) already suggested a
typology of interdependence between policy actors in terms of the relative
importance of the resource and the substitutability of the source. For example,
when an actor has highly important resources that have low substitutability, the
other policy actors show high dependence on the actor.
More recently, Fenger and Klok (2001) focused on interdependency and
belief congruence among policy actors in explaining coalition behavior in the
policy process (see Table 3.3). They argued, for example, that strong coordination
occurs when policy actors share congruent beliefs and symbiotic relationships. In

contrast, divergent beliefs and competitive relationship portend possible strong
conflict between policy actors.
Table 3.3
Coalition Behavior as a Result of Interdependency and Belief Congruence
Interdependency Beliefs
Congruent Indifferent Divergent
Symbiotic (1) Strong coordination (2) Coalitions of convenience (3) Unstable conflict, depoliticization, learning
Independent (4) Weak coordination (5) No coalitions (6) Weak conflict
Competitive (7) Coalition with severe collective (8) Weak conflict (9) Strong conflict
action problems
Source: Fenger and Klok, 2001.
Similarly, Lubell et al. (2002) suggested that collective action is influenced
by problem severity, institutional opportunities (such as the degree of integration of
agencies), political incentives, institutional support (such as federal aid), and
characteristics of the actors. Also, Schneider et al. (2003, p. 152) argued the
emergence of trust and norms of cooperation based on repeated interactions can
foster collective action in policy communities even in the presence of conflicting
values and beliefs. Focusing on the relationship between Congress and non-
Congressional government bureaucracies, Ripley and Franklin (1984) suggested

some conditions for cooperation or conflict between them. For example, they
argued that a high degree of compatibility between key members of Congress and
the governmental agencies promotes cooperation.
Another central aspect in collective action among policy actors comes from
the competition among bureaucratic agencies. In most cases, bureaucratic agencies
are core policy actors but their interaction modes are not pre-determined. Although
the agencies share the same umbrella of government, the coordination of
bureaucratic actors is not always guaranteed. Instead, competition among the
agencies is often observed, caused either by legislators intent to control
bureaucracy or by the results of bureaucratic agencies efforts to keep their
autonomy and territories (Downs, 1967; Miller & Moe, 1983). In particular, the
concepts of allocational rivalries and functional rivalries reflect the motives and
targets of the competition (Nicholson-Crotty, 2005). That is, allocational rivals
compete with each other despite their sharing of policy goals because resources are
often not sufficient to satisfy all the demands from policy actors. Functional rivals
are different from allocational rivals in that they pursue different policy solutions
based on their own value systems.
Klijn and Koppenjan (2000) stressed network management as a major factor
contributing to cooperation among policy actors in the policy process. They
suggested two types of network management: process management and network
constitution. The former is mainly about developing relationship among policy

Full Text
2009 by Young-Jung Kim
All rights reserved.

Kim, Young-Jung (Ph.D., Public Affairs)
Mapping the Policy Networks: A Case Study of the Korean Foreign Labor Policy
Thesis directed by Professor Peter deLeon
Although recent advances in theorizing the international migration policy
are apparently remarkable, the greatest weakpoint of the literature on international
migration policy lies in its lack of consideration for the policy process, in which
policy actors make real interactions and produce ultimate policy outcomes. In this
sense, this thesis tried to add value to the extant international migration policy
studies by focusing on the process of the Korean foreign labor policy; particularly,
this research attended to the policy networks in the policy process.
In this thesis, the following four specific themes were spotlighted and
intensively examined with 25 hypotheses: (1) the glue of policy coalitions; (2)
policy actors collective action; (3) trust relationships in a policy network; and (4)
policy actors influence. In approaching those themes, the research adopted
analytical methods from both social network analysis and traditional quantitative
analysis tools, such as multiple regression and correlation analysis.
Major findings of this research include the following: (1) all the variables
of the proximity of belief system between policy actors (PBS), the proximity of
policy interest between policy actors (PPI), and the proximity of policy preference

I dedicate this thesis to my mother, Yun-Im Yu, for her inexhaustible affection and
her constant reminders that life can be better off with perseverance and resolve. I
also dedicate this to my wife, In-Moon Kim, and my little princesses, Mi-Sung and
Min-Jung, for their continuous support and encouragement during the long time
required for completion of this thesis.

1. INTRODUCTION....................................................1
Knowledge in the International Migration Policy Process.1
Knowledge of the International Migration Policy Process.3
Research Purposes and Questions................................6
Structure of the Dissertation.................................11
Analytical Framework..........................................17
Basic Legal Systems Regulating Foreign Workers................18
Major Changes in the Korean Foreign Labor Policy..............20
The First Round (1989 1991): Adoption of the Industrial
Training System for Foreigners.........................20
The Second Round (1994 1998): Emergence of the EPSF as an
alternative to the ITSF................................29

Case Study Analysis........................................108
Analysis Targets...........................................113
Research Framework................................................125
Operationalization of Variables............................127
Research Hypotheses...............................................136
The Glue of Policy Coalitions..............................136
Collective Action in a Policy Coalition....................140
Trust in a Policy Network..................................143
Policy Actors Influence in a Policy Network..............145
Data Collection...................................................148
Archival Review............................................149
Data Analysis.....................................................159
Network Analysis...........................................159
Multiple Regression........................................166
5. RESEARCH FINDINGS..................................................169
Descriptive Analysis Results......................................170
Major Characteristics of Respondents.......................171
Policy Actors Institutional Attributes....................177

EPSF DURING THE YEARS OF 2002 AND 2003.......258
B. SURVEY QUESTIONNAIRE.........................261